Determining the Economic Risk/Return Parameters

DETERMINING THE ECONOMIC RISK/RETURN PARAMETERS FOR DEVELOPING A MARKET FOR ECOSYSTEM GOODS AND SERVICES FOLLOWING THE RESTORATION OF NATURAL CAPITAL: A SYSTEM DYNAMICS APPROACH

Volume 1: Main Report

James Blignaut, Martin de Wit, Sue Milton, Karen Esler, David le Maitre, Steve Mitchell and Doug Crookes (Editors)

Report to the

Water Research Commission

and the

Working for Water Programme Department of Environmental Affairs

WRC Report No. 1803/1/13 ISBN 978-1-4312-0435-9

JULY 2013

OBTAINABLE FROM

Water Research Commission Private Bag X03 Gezina, 0031 [email protected] or download from www.wrc.org.za

This report forms part of a series of two reports. The other report is Determining the economic risk/return parameters for developing a market for ecosystem goods and services following the restoration of natural capital: A system dynamics approach. Volume 2: Policy Briefs. (WRC Report no 1803/2/13)

DISCLAIMER This report has been reviewed by the Water Research Commission (WRC) and approved for publication. Approval odes not signify that the contents necessarily reflect the views and policies of the WRC nor does mention of trade names or commercial products constitute endorsement or recommendation for use.

Executive summary

Project overview

The degradation of ecosystems, mainly driven by human-induced land transformation, has reached unprecedented levels at a global scale and has been well documented (Wake and Vredenburg 2008; Estes et al. 2011; TEEB 2011). Various studies have indicated the unsustainability of the prevailing situation and the need for a rapid, sustained and large-scale intervention (sic. ecological restoration) to prevent the failure of ecological systems and the socio-economic loss of welfare associated with it (MA 2005; Halpern et al. 2008; Rockström et al. 2009, Bulchart et al. 2010; CBD 2010; LPR 2012).

Most of the cost of this degradation, or the loss of natural capital, has been borne by the environment and society, for both the current and future generations, an outcome known in economics as externalisation. This has happened because there is no measure of the current cost or the benefit that ecological restoration would bring to society because there is no market to record these exchanges.

South Africa has developed a proud history of addressing concerns about environmental degradation by means of restoration driven through the package of Natural Resource Management programmes – which have affectionately become known as the “Working for” projects. In addition to these restoration interventions, South Africa also has a series of national environmental policies and legislation (e.g. NWA, NEMA, CARA) that make provision for mandatory restoration projects, such as mining rehabilitation. None of these efforts, despite their obvious value and worthiness, effectively internalise the costs and the benefits of restoration activities. As symptomatic responses to an ongoing problem, they fail to change the economic drivers that generate the need for restoration. This happens largely because the cost of degradation and the need for, and value of, restoration are not explicitly considered.

This study focused on developing an evidence base for the use of economic tools/instruments in the decision-making process about the restoration. We have neither investigated the need, nor the moral and/or intrinsic reasons, for restoration. We applied the conventional return/risk economic decision-making framework to eight existing restoration projects to evaluate whether such a decision- making framework could be applied to restoration over a range of environmental conditions and in different contexts. By making both the cost and the benefits of restoration explicit, we aimed to illustrate the potential for the development of markets for ecosystem goods and services (offered by restoration). Our underlying assumption was that by changing market signals, market participants will adjust their behaviour.

Background

Through the Natural Resource Management programmes – mainly Working for Water, Woodlands, Wetlands and Fire projects, and various other restoration projects such as the mandatory restoration and rehabilitation of road servitudes and mine-dumps, South Africa has established itself as a country which has started to invest in the restoration of natural capital (RNC). This is especially true among developing nations where South Africa is being used as an example and leader. However, despite this iii rich history of restoration, no meta-analysis (a high-level, cross-sectional analysis) has been done to assess restoration’s ecological, hydrological, and economic impacts across a range of contrasting sites and contexts. This study aimed to rectify this obvious deficiency.

The specific objectives of this study were to conduct ecological, hydrological, and socio-economic assessments to determine the impact of restoration at eight existing ecologically and socio- economically different restoration sites by comparing them with degraded or un-restored areas in close proximity. The outputs from these studies were used to develop an integrated system dynamics model on the likely impact of restoration on the ecology, hydrology and economy of notably agriculture. This model was specifically focused on internalising the economic (societal) costs and benefits of restoration and to apply an economic decision-making rationale to the results in an effort to make the societal benefit of restoration explicit.

Restoration impacts positively on flows of a suite of ecosystem goods and services and therefore, we suggest, on the economy. While some evidence exists of the ecological and hydrological implications of restoration for individual projects, the links between these restoration activities and the economy across various spatial scales and biomes have not been established. There is also no clear understanding of how the benefits before and after restoration might affect agriculture through improved returns from terrestrial ecosystems. These benefits generally are believed to be very real and significant but they are not well understood. This study endeavoured to provide these links. In effect, restoration is no different from the capital expenditure on any project and the return to the land. The value of environmental services emanating from the ensuing flows (as a result of the capital expenditure) is the annual stream of benefits delivered at an annual maintenance and operation cost. This is not unlike any other investment that does have an upfront capital component with regular/annual operational and managerial cost, but that yields an ongoing stream of benefits in the form of products or services being sold.

Although South Africa has a proud history of RNC, a meta-analysis of the ecological, hydrological, and economic impacts of restoration across a range of contrasting sites and contexts is lacking. This study aimed to address this deficiency and to determine the tangible (societal) contributions of restoration in terms of economic parameters. We focus on existing restoration sites and monitor and evaluate the ecological, hydrological and socio-economic impacts restoration had at those different sites. We used a carefully selected set of ecological, hydrological and socio-economic parameters to test the following hypothesis at eight different sites:

RNC improves water flow and water quality, land productivity, in some instances sequesters more carbon, and, in general, improves both the socio-economic value of the land in and the surroundings of the restoration site as well as the agricultural potential of the land.

The distribution of the sites is shown in Figure a below, with a summary of the site characteristics provided in Table a.

Figure a Geographical distribution of case studies Table a Description of restoration study/project sites

Site Biomes Climatic MAP Ownership Size Extent of Zone (km2) Degradation 1 Succulent Arid 160 Private 26 Severely degraded: Restoration following Karoo open-pit surface mining 2 Nama Karoo Arid 239 Public/ 8 Degraded: Clearing of invasive alien Private plants 3 Succulent Arid 242 Private 1,762 Severely degraded: Restoration following Karoo overgrazing ostriches 4 Savanna Semi-Arid 400 Private 9,249 Degraded: Bush-thinning (and combating bush encroachment) 5 Fynbos Semi-Arid 478 Public/ 548 Degraded: Clearing of invasive alien Private plants 6 Fynbos Semi-Arid 650 Private 46 Degraded: Clearing of invasive alien plants in the riparian ecosystem 7 Grassland Temperate 900 Communal 1 Severely degraded: Restoration of a communal grassland system following overgrazing 8 Forest/ Temperate 1275 Public/ 32 Degraded: Removal of exotic plantation Savanna Private forestry MAP = Mean annual precipitation; size refers to the size of the study site from an economic perspective

Objectives

The objectives of the study were:

1. To conduct ecological assessments to determine the impacts of restoration at eight ecologically and economically different sites in comparison to degraded or unrestored areas in close proximity to the selected sites. 2. To conduct hydrological assessments to determine the impacts of restoration at eight ecologically and socio-economically different restoration sites in comparison to degraded or unrestored areas in close proximity to the eight selected sites. 3. To conduct economic and socio-economic assessments to determine the impact of restoration at eight socio-ecologically and socio-economically different restoration sites in comparison to degraded or unrestored areas in close proximity to the eight selected sites. 4. To investigate the impact of restoration on sustainable rural employment and payment for environmental services. 5. To develop a user-friendly model that could be used in future to model the likely impact of restoration of the ecology, hydrology and economy – notably agriculture. 6. To prepare a meta-analysis as a synthesis of all the studies above based on the outcomes of the research.

To meet these objectives the study was divided into 16 deliverables, each of which is available in report format from the WRC. This report is deliverable 16, the final synthesis report. In addition to producing the 16 deliverables, the study has generated a range of further outputs, namely:

• Peer-reviewed academic papers (8) • Policy papers (5) – submitted in association with TIPS (Trade & Industrial Policy Strategies) • Posters (15) • Articles in popular magazines (7) • Press releases (2) • Oral conference presentations (10) • Student colloquiums held (8) • International short courses presented (1) • Various meetings with stakeholders

Approach

This study used a carefully selected set/suite of ecological and socio-economic parameters to test the hypothesis (see Background). The overarching systems study and those at each of the eight sites were based on the generic research design described below. The research team identified eight existing restoration sites that were well established and that had significant/sufficient supporting data and allocated the study sites to student teams. Each student in restoration ecology carried out an ecological assessment of the impact of restoration on ecosystem function in his/her site. Together they covered all eight sites, asking some of the same questions and doing similar types of work. The same applied to vi the hydrological and economic assessments. The basic field study design included a comparison of 4 types of (sufficiently similar) sites, such as:

• baseline sites comprising natural vegetation in good condition, • pre-treatment (or untreated) invaded or degraded, • cleared of livestock, but not rehabilitated, and • rehabilitated (including alien removal, erosion control, species re-establishment).

In the sites where hydrological assessments were conducted, the studies focused on how invasions and clearing have affected the hydrological functioning of a site:

• One study derived a basic catchment water balance based on the difference between the rainfall and the estimated transpiration and interception losses for sub-catchment units based on data on the vegetation structure and evaporation. • The other studies assessed the relative quantities of water that become overland flow versus that infiltrating the soil using combinations of runoff plots, infiltrometers and measures of the soils porosity and proneness to form mineral crusts (e.g. Mills & Fey 2004a, b). • The site-based studies collected a number of indices of site hydrological function based primarily on the approaches developed by David Tongway and others; these used various indices to quantify the nature of the soil surface, vegetation (e.g. basal cover) and indicators of surface water flow and sediment transport. These indices were developed largely from studies of patchy semi-arid vegetation and reflect the basic hypothesis that “pristine” landscapes retain resources while degraded ones leak resources (Ludwig et al. 1997 and subsequent papers; Belnap et al. 2008). These basic hypotheses and the indices were tested for local applications incorporating local experience and knowledge (see Le Maitre et al. 2007 on the Little Karoo).

The methods were standardised across the different studies as far as possible so that all the hydrological function indices and vegetation measurements were collected on all the sites. The actual measurements of infiltration-related variables varied from site to site depending on the conditions, but a minimum set was collected by all the site-based studies. Each hydrological study was run in close conjunction with an ecological study and contributed to, and used, the historical and ecological information collected for each site. The students involved in the hydrological studies were required to review the available methods and to obtain expert opinions on their applicability before the final set of measurements was chosen. The hydrological studies were aimed at testing a range of relatively straightforward indices of site water balance and hydrological function that can be used by land managers in a range of vegetation types and conditions. There are existing soil erosion models that could have been used but there are still many issues to be addressed (Boardman 2006) so erosion modelling was not envisaged. However, soil stability (which is a measure of erosion potential) was assessed using landscape function indices.

There were two components to the site-based ecological assessments, namely:

• a desktop review of ecological information for a particular site, which included a thorough review of the history of the site, the ecological goals and the future land-use for that particular intervention; and

vii

• a field component focusing on understanding the impact that the restoration activity has had on ecosystem diversity, structure and function.

Each ecology study used a combination of line transects and/or plots to assess the following variables where applicable:

• cover of alien and indigenous plant species • species richness of alien and indigenous plant species • vegetation structure • resilience (changes in cover between seasons) • grazing capacity (ha/LSU) (where LSU = Large Stock Unit) • soil structure • assess the change in ecosystem function before and after restoration.

This information was used to determine the impact restoration has had on the flow and generation of ecosystem goods and services, with a particular reference to its value for agriculture.

The economic studies focused on the same eight sites and used the information generated by the ecological and hydrological studies, in combination with local level surveys/questionnaires and/or group sessions, to determine the socio-economic value of RNC in each case. This was done to determine the economic value of the site before and after restoration, or in comparison with a control site/area. The socio-economic assessments were conducted using surveys of the local interested and affected parties. Thus in most cases both an economic and financial cost-benefit analysis was done to determine the economic viability and feasibility of restoration. This portion of the study explicitly looked at the possibility and/or feasibility of embarking on a trade or payment system for ecosystem services and the benefit to agriculture restoration is providing.

The study outcomes were synthesised and analysed through the use of a System Dynamic model using VENSIM software. System dynamics (SD) models simulate how complex systems change over time, and provide a means of capturing realistic dynamic behaviour between variables, including feedback loops, delays and nonlinearities. The system dynamics approach is consistent with traditional economic approaches to modelling dynamic phenomena (Smith and van Ackere 2002). However, the advantage of SD modelling is that it can deal with disequilibrium conditions, as well as provide a realistic portrayal of the processes involved in decision-making (Sterman, 1987):

The purpose of simulation models is to mimic the real system so that its behaviour can be anticipated or changed. Behavioural simulation models must therefore portray decision-making behaviour as it is, and not as it might be if decision-makers were omniscient optimizers. The decision-making heuristics and strategies people use, including their limitations and errors, must be modelled.

This approach is appropriate for modelling complex interactions between social, economic and environmental components.

This study was student-driven and involved several University departments and institutions (Table b).

viii

Table b List of students involved in the study by site and discipline

Registration Qualification Student University Site Discipline Year (degree) Agulhas and Helanya Vlok 2009 MComm Economics SU Economics Oudtshoorn MSc Conservation Ecology with Marco Pauw 2009 SU Namaqualand Ecology hydrological aspects MSc Conservation Ecology with Megan Nowell 2009 SU Agulhas Ecology hydrological aspects Thabisisani MSc Conservation Ecology with 2009 SU Beaufort West Ndhlovu Ecology hydrological aspects MSc Conservation Ecology with Petra de Abreu 2009 UCT Oudtshoorn Biology hydrological aspects MSc Agricultural Oudtshoorn and Worship Mugido 2009 SU Economics Economics Namaqualand MSc Conservation Ecology with Alanna Rebelo 2010 SU Kromme river Ecology hydrological aspects Kromme river Katie Gull 2010 MSc Economics UCT Economics and Lephalale Ecology with Jacques Cloete 2010 MSc Natuurlewe UFS Lephalale hydrological aspects MSc Conservation UCT (and Ecology with Dane Marx 2010 Drakensberg Biology UKZN) hydrological aspects Drakensberg, Sand river and Douglas Crookes 2009 PhD SU Economics systems-dynamic modelling SU = Stellenbosch university; UCT = University of Cape Town; UFS = University of Free State; UKZN = University of KwaZulu-Natal

Results

This study considered economic return and project risk as its main determinants of economic success. This section addresses project returns and then project risks.

The results with respect to economic returns are provided in Table c. Most net present values (NPVs) are positive and range between R300 million and R620 million after 50 years, once all ecosystem benefits are taken into consideration (cultivated, replenishable and renewable). The size of the restoration project, however, is very important. When the results are expressed per unit area, the more arid areas typically generate relatively low values compared with the higher values for more humid areas (Figure b). This assessment, however, ignores risk and uncertainty.

Table c Summary of present values for the different restoration sites (Rand million)

Area PVs of restoration (R million) Total (ha) Benefits – Benefits – Benefits – NPV

Cost Cultivated Replenishable Renewable

Namaqualand 2,619 -93.4 -3.1 0.0 0.7 -95.8

Beaufort West 781 -2.3 0.2 3.9 1.3 3.0

Oudtshoorn 176,216 0.0 30.8 0.0 15.4 46.2

Lephalale 924,920 -5,523.1 5,090.1 0.0 910.2 477.3

Agulhas 54,755 -12.1 0.0 419.4 -8.6 398.6

Kromme 4,640 -98.9 34.0 358.6 0.0 293.7

Kromme – no agriculture 4,640 -98.9 0.0 240.7 0.0 141.8

Drakensberg 90 -2.0 0.0 3.9 0.3 2.2

Sand river 3,216 -24.3 368.7 304.3 -35.9 612.8

250000

200000

150000 B1 Cultivated 100000 B2 Replenishable B3 Renewable 50000 Rand/hectare

0 Nam BW Ou Lp Ag Kr Dg S -50000

Figure b Cumulative NPV (Rand/hectare) for each of the different study sites Notes: sites arranged in decreasing order of aridity. B1 indicates the present value (PV) of restoration costs plus PV cultivated benefits; B2 indicates B1 plus replenishable natural capital PV, and B3 is B2 plus renewable PV (i.e. total NPV). Study sites: Nam=Namaqualand; BW=Beaufort West; Ou=Oudtshoorn; Lp=Lephalale; Ag=Agulhas; Kr=Kromme; Dg=Drakensberg; S=Sand

We turn now to evaluating the economic risk inherent to a restoration project. One method to address risk and uncertainty is through portfolio mapping. A conventional way to display portfolio mapping is to focus on the relationship between return (as expressed in terms of NPV) and risk (as expressed in terms of the probability of success of a project), also taking project cost into consideration. The standard and most commonly used portfolio map is the risk reward bubble plot (Figure c), with the size of the bubble indicating resources committed to it. This provides a way to compare projects by considering a range of factors (reward or payoff, probability of success, and cost). These projects are independent of each other, so the total resource cost will not add up to the budget. Furthermore, although some projects show a negative NPV, this is only because the project costs are compared with one ecosystem good at a time, rather than the entire range of ecosystem goods and services (EGS) that are assessed for the project as a whole.

Results indicate that in semi-arid South Africa, restoration projects yielding water services are the ‘pearl’ projects, with high likelihoods of success and high payoffs. Restoration projects yielding grazing and crop services are mostly the ‘bread and butter’ projects, ones which are likely to succeed but yield low (negative?) rewards. There is one ‘white elephant’, the Namaqualand mining project, with large resources committed to it, proportionally little reward and low probability of success in terms of restoration outcomes. As will be noted later, however, despite the cost the restored area’s level of ecosystem function is still below that of the undamaged area, indicating an unmitigated loss despite restoration. The project, however, is a legal requirement placed on the mining company as part of its licence to operate and which changes the context and, therefore, requires a different type of evaluation. In this case we evaluated the project in terms of benefits and costs of ecological restoration to make it comparable across sites, and we did not include the added benefits to the mining company of legal compliance. Lephalale (grazing) is a potential oyster, with untested and therefore uncertain long-term benefits from restoration. Fairly low levels of resources are committed to this activity.

1.2

0.8

0.6 Grazing Water -10000 -50000.4 0 5000 10000 15000 20000 Crop 0.2

-0.2

Figure c Portfolio map for different ecosystem services (bubble size indicates resources committed to it). Study sites: Ag=Agulhas; BW=Beaufort West; D=Drakensberg;

Ka=Kromme (agriculture); Kna=Kromme (no agric); Lp=Lephalale; N=Namaqualand; S=Sand

There is, however, no individual measure of risk (e.g. success probability, standard deviation, CV) that is sufficient for selecting and classifying projects. A combination of measures provides an improved means of selection. This is then used to inform a portfolio mapping exercise, in order to classify and select restoration projects (Table d). A summary of the classification of projects suggests that those with the highest potential payoffs (the pearl projects) are the water projects, in other words those projects where downstream water consumers benefit from the restoration project. The Agulhas, Beaufort West, Kromme and Sand study sites are all examples of this.

Table d Summary of projects classified by type Oyster Pearl Bread and Butter White elephant High risk projects Projects with high Essential projects Projects which are with uncertain likelihood of that enterprises preferable to Description merits success cannot do without avoid Agulhas, Beaufort Drakensberg; West, Kromme Kromme (no (with agriculture), Water projects agriculture) Sabie Sand Agulhas, Kromme Crop projects Sabie Sand (with agriculture) Beaufort West, Drakensberg, Oudtshoorn Kromme (with Grazing projects Lephalale (passive only) agriculture) Namaqua Sands

Contribution of the research

This study is, for the most part, an economic one, albeit an investigation into the economics of the restoration of natural capital or restoration economics. Restoration is not only an ecological or hydrological activity or focus of investigation; sound ecological principles and information are needed but, to succeed, restoration has to be possible economically and practically as well (Hobbs and Harris, 2001). This study does not seek to address or provide answers to hard-core ecological and hydrological questions with respect to where ecological and hydrological thresholds and tipping-points are, nor is the intention to provide an ecological/hydrological motivation for restoration.

Here we work within the framework of eight existing restoration projects and therefore, a priori, accept the need for restoration. However, these projects lacked the context of an operating market for restoration and, prior to this investigation, had not considered the potential improvements in ecosystem goods and services rendered by restoration. We therefore reflect back and ask the question: Can markets assist by providing support for restoration and, if so, under which conditions? To do this we studied both the bio-physical and socio-economic dimensions of the restoration conducted.

Our study provided some unique perspectives in that it was reflective, and multi-disciplinary. Studies rarely look back and evaluate the impact of projects and even more rarely do they do this in a multi-disciplinary way covering various study sites. Our proposal for an integrated approach to the ecology, hydrology and economics of restoration uses the classical economic calculus of costs and benefits as a starting point for evaluating restoration interventions, while building on empirical work in

xii the fields of ecology and hydrology. The concepts associated with complex and dynamic systems are therefore taken as a given and seen as an inherent part of the study; although we conducted sensitivity analysis, this was not added on at the end in an adjunct manner. We selected this approach since ecological systems have a number of individual components that interact in non-linear ways over a multiplicity of scales, while being heterogeneous across space (Wu 2002). To effectively manage (restoration as one option) such systems we required at least an understanding of the properties and dynamics of such systems (see for example Maler, 2000).

We demonstrate that standard economic evaluation methods have limited ability to take system dynamics into account and could, therefore, easily misdirect or reject investment in restoration, potentially leading to disastrous social and ecological consequences. The alternative often presented, namely where the concept of critical ecosystems or ecological thresholds needs to be defined a priori (Farley & Brown Gaddis, 2007) is also not followed in this research. Ecological thresholds are in fact very hard to determine without actually crossing them (Suding and Gross, 2006).

We propose a systems dynamic (SD) approach, building on what the economics profession and the literature on complex and dynamic systems already have to offer, but applying this to restoration. Contrary to some suggestions, there is no need to abandon conventional economic cost-benefit evaluation tools when asking under which conditions markets could make a contribution to restoration. These conventional tools, when enriched with an understanding of system properties and their dynamics, can be used fruitfully to shape decision-making regarding restoration priorities; furthermore, they can assist in development of markets and/or payment mechanisms for ecosystem services. Such an approach moves beyond standard static economic evaluation approaches as discussed, for example, by Figueroa (2007) and provides a novel way to move beyond the contested use of an exogenously determined discount rate as a single variable to linearly reflect the value of costs and benefits over time (see Mills et al. 2007, Holmes et al. 2007 for an application of cost-benefit analysis with exogenous discount rates in the context of restoration). By using a SD approach, it is also feasible to simulate repeated random sampling of uncertain inputs, and therefore to generate a measure of risks in restoration investment decisions. We demonstrate that the ensuing risk/reward outcomes provide a far more nuanced and thorough way of evaluating any project, including restoration projects, than the conventional net present value (NPV) outcomes favoured in most natural resource economic evaluation projects.

The benefit of an SD approach is that decision-making about payments for ecosystem services from restored ecosystems are now driven by the known or expected changes in properties of that system. This is quite different and much more nuanced and sophisticated than the application of exogenously determined discount rates. Discount rates are usually used in a static framework of costs and benefits over time, often to account for much more than what they were originally intended for, namely to act as a proxy for people’s preference of holding money over time. Although we used a discount rate to reflect the value of money over time, it had no bearing on the relative ranking of projects in terms of whether markets can or cannot contribute to restoration. That ranking was decided on bio-physical and socio-economic fundamentals inherent in each project. The market- development decision-making priority list is therefore discount rate neutral. This is a further significant departure from conventional methods.

Using the proposed decision-making framework with respect to the development of markets/payment systems for ecosystem goods and services following restoration can now be taken xiii against the backdrop of the risk involved in achieving such rewards or benefits. Neither SD approaches nor risk quantification by themselves are new, but applications to existing and on-going restoration projects are novel. This study hopes to contribute to the science and practice of restoration through such an evidence-based approach to integrating economic evaluation and ecosysttems dynamics.

On a final note. Since this study did not seek to provide a motivation for restoration, but only sought to identify under which conditions markets could contribute to restoration, we do not suggest that only monetary values are of importance within the larger restoration deccision-making picture. Those restoration options that have high risk/low reward outcomes over time should not necessarily be abandoned; we only suggest that markets are ill-equipped to assist in resstoration under such conditions. We acknowledge that there may be a suite of other drivers for doing restoration, such as legislation on mining for example, where restoration needs to be conducted according to legal requirements and also socio-economic considerations like job creation (see Figure d for a diagrammatic interpretation).

Figure d When to use markets to assist in restoration activities

xiv

Conclusions

The analysis of projects using portfolio mapping suggests that this approach, coupled with risk analysis and system dynamics modelling, is able to provide a means of selecting and prioritising restoration projects based on economic parameters.

Technical / operational recommendations

Use of conventional economic evaluation tools: This study showed that conventional economic evaluation tools may be used for the valuation of the restoration of natural capital.

We recommend that conventional economic evaluation tools are used for future valuations, but it is necessary to recognise the dynamic nature of ecosystems with their inherent risks and uncertainty and to accommodate these when using the tools.

Costs of restoration: The high cost of restoration, together with the lag before benefits are realised, makes the benefit cost ratio (BCR) unprofitable for private entities in most cases because the benefits are restricted to those yielded on site. But restoring ecological infrastructure in one place may lead to improved delivery of ecosystem services in another place. An example of this is the water that is made available through restoration for a city, such as Port Elizabeth, which depends on the transfers of that water. The cost of delivering this water in urban and peri-urban areas was found to be on a par with what users currently are paying in the drought-prone areas of the country. In most of the case studies it was the value of water that made restoration economically viable which is appropriate given that South Africa has limited water resources and the current costs of water do not reflect the real costs of delivering it. This failure to internalise the real costs is one of the main factors leading to the inefficient use and wastage of water and limiting efforts to recycle it.

One of the greatest barriers to realising the value of restoration to society is the high upfront cost of restoration. This problem is not unique to water – it is characteristic of all infrastructure (manufactured or man-made) investments which are a means to an end and aimed at delivering a vital public good or service rather than a purely private one.

We recommended that restoration of natural capital is recognised as a water supply option and needs to be seen as an investment in ecological infrastructure that yields, among others, water services.

Restoration potential assessment protocol: Restoring natural capital seeks to rebuild ecosystem resilience, thereby replenishing natural capital stocks and improving the flow of goods and services to society. But the governance of environmental issues is fragmented such that interventions tend to be department specific when cross-sectoral interventions would be more effective if not essential. We designed a framework to guide decision-making on the restoration of natural capital. This framework address three main components (a) issues of governance focused on how to make this work; (b) the environmental factors that can determine the options and the outcomes; and (c) who will benefit from this and how well matched the benefits and their needs are.

We recommend that the framework for restoration (see Fig 5.1) be used to guide decision making by the project proponents, stakeholders and funders. Use of this framework should also take into account the ecological, social and economic risks as well as the potential returns on those investments.

Restoration as an insurance against risk: There is a pressing need to supplement financial insurance with nature-based insurance policies which provide a hedge against adverse impacts on natural capital. The similarity between financial insurance and ecosystem-based approaches has been recognised (Van Oosterzee et al. 2012). Ecosystem services themselves, however, have no recognised hedge mechanism. We proposed adapting van Oosterzee et al. (2012) approach to hedging against risk in a REDD (Reduction in Emissions through Degradation and Deforestation) programme, to provide a mechanism hedged against a failure or malfunctioning of natural systems, and thus the reduced or non- delivery of ecosystem goods and services, as a result of degradation and a loss in quality. We suggest that restoring the functionality of degraded ecosystems will offer a bio-physical, system-wide, insurance against the effects of adverse events. Ecological restoration activities to restore resilience are therefore likely to safeguard against the loss of essential services, as well as against the consequent loss of manufactured, cultivated and human capital.

We recommend that, where shown to be viable, degraded ecosystems be restored. In addition, recognising that the premium paid to keep ecosystems in good condition is lower than that paid to restore those that have been damaged, we recommend that a long-term view of ecosystem management is taken. This entails conserving what we have rather than looking for short-term gains which will permanently(?) reduce the stream of goods and services.

Restoration and the environmental reserve: Modelling of the restoration of natural vegetation in the Sand River catchment will release sufficient water to meet the requirement of the ecological reserve. In the case study on the Sand River catchment, the downstream beneficiaries within South Africa are the conservation areas. However, a number of questions arise around the payment for the restoration and these are detailed in Section 5.4.

We recommend that, where water is needed to meet the requirement of the ecological reserve, the option of restorating the natural capital be investigated. In doing this, consideration should be given to the questions raised in Section 5.4.

The cost to society of unmitigated damage to the environment: Environmental legislation provides a mechanism to protect society from the negative effects of development on natural environments and the services that they supply, and if such damages do occur, to manage those. Government authorisations are required before environmentally damaging activities may occur, with the ‘Record of Decision’ (RoD), outlining the criteria by which a destructive activity may proceed. The regulations are very specific about the need to quantify potential damages, losses and risks, and to develop environmental management plans (EMPs) that specify how damage and costs to society should be avoided, minimised and mitigated. An entity embarking on an environmentally destructive activity might also be required to restore the environment following such an activity. In the case of mining, for instance, government may set monitoring and auditing protocols to ensure compliance with restoration targets set in the EMP, before they would issue a mining closure certificate.

xvi

We recommended that environmental management plans be drawn up in a way that, as far as possible, ensure that there is no unmitigated damage to the environment or to the social- ecological systems reliant on the benefits from the environments. However, where this is not possible, a suitable offset should be required by legislation.

Current restoration initiatives: Restoration is currently being undertaken in South Africa, with much of it being conducted through the Natural Resource Management projects of the Expanded Public Works Programme. The effectiveness and efficiency of these interventions could be improved by taking the restoration to an appropriate end point using the methods for assessing the success of each intervention used within this project, while bearing in mind its limited duration.

We recommend that the methods advocated in this report, to set an appropriate end point and to measure the success of the restoration intervention, should be applied more widely through other projects, potentially improving their success rate in terms of evaluating risk and reward.

Restoration should be anchored in the local economy: In, among others, the Kromme River case study, the research team spent time explaining the outcomes to the farming community after the main project was over. This has ensured that there is a high level of understanding of the purpose and benefits for restoration work.

We recommend that owners or users of land on which restoration interventions are being undertaken are conversant with, and supportive of, the restoration in order to motivate them to support and maintain the work. This does, however, require buy-in by the authorities.

We recommend that the beneficiaries of restoration contribute, either in part or in full, towards the costs of the restoration work.

Temporal and spatial dynamics matter (e.g. initial lag, time to full effect): In each case there is a time lag between the restoration intervention and the realisation of the benefits of the interventions.

We recommend that restoration interventions should be planned and executed knowing who will be the beneficiaries of the activities, and these beneficiaries should be willing to pay for at least part of the benefits of the restoration process. This is particularly important where occupiers of the land may be expected to play a part in maintaining the interventions while receiving only part of the benefit from the restoration or during the lag period before the benefits of the restoration are realised.

Return on investment: The studies have provided methods to assess the level of risk and the return on funds invested in restoration. There is a clear relationship involving the trade-off between risk and return and where investors are likely to go first.

xvii

We recommend that the costs of restoration should be apportioned to the beneficiaries, taking into account the socio-economic and risk profile and return on investment.

Technical – environmental science-based aspects: Only some ecosystem services are traded in the market and so the economic value of some services is often not recognised.

Recommendations:

• Create capacity to understand the national need for restoration and also what can or needs to be done at both the national and the local levels;

• Develop a greater understanding of the concept of natural or ecological capital (or infrastructure) and how this contributes to the socio-economy;

• Natural capital needs to be valued and these values added/injected into the budget along with man-made or manufactured capital;

• When considering how interventions may be funded, factors to consider are the opportunities for the payment for ecosystem services and the possible time lag between intervention and benefits. These will influence who may be approached to anchor the project;

• Planning of restoration interventions should take account of the position in the landscape in relation to the kinds of services and benefits being addressed.

Governance: National policies and legislation (e.g. NWA, NEMA, CARA) provide the framework within which restoration projects are supported, but planning and decisions for individual restoration projects should be based on a participatory approach.

Recommendations:

• Ways need to be found to ensure that land owners / occupants are committed to work with the restoration and maintain it as necessary;

• Government stimulation may be required in the absence of existing markets;

• Policy and legal goals and imperatives, where these apply, provide the overarching framework within which restoration will be carried out;

• Capacity / capability for the implementation of legislation should be developed where this is weak.

Institutional: Where there is private or government ownership of land or resources the institutional arrangements within which restoration projects are implemented are clear but, where tenure is communal or open access, then the institutional arrangements will need to be clarified and supported.

Recommendations:

• The costs and benefits (both public and private) of the restoration intervention are defined and understood by all parties involved;

xviii

• Where at least some of the beneficiaries of a restoration intervention are off-site it is recommended that the relationship between the beneficiaries and the occupants of the land is formalised in a way that will protect the interests of all parties;

• Residents of the site and beneficiaries should all work towards the same goal;

• Make provision for, and/or develop the means and the mechanisms whereby local communities and/or conservation agencies can raise the funds for restoration work through a Green Fund or the like.

Future research

The following is a non-prioritised list of questions for future research that stems both directly and indirectly from the research conducted.

Legislation and Policy: There is good legislation for protecting ecosystem services. Of particular interest to restoration are the environmental management plans (EMP) that are required when an activity will damage the environment. This project identified the following problem with the EMP for the mining in Namaqualand which we believe is probably widespread: even though the mining in Namaqualand has met the terms and conditions of the EMP, the ecosystem services delivered by the restored area are inferior to those from the undisturbed area, resulting in a net loss to society. Addressing this deficiency will require two measures: firstly, ensuring that there is no net loss to society from the activity (in this case it would have included restoration of the palatable plants and the shrub species needed to provide and sustain the grazing capacity); and, secondly, ensuring that there is sufficient flexibility to allow for the setting of intermediate targets, with the final target, especially in arid areas, being long term.

Research question. How can the EMP requirements be changed to accommodate the greater flexibility required to ensure that there is no unmitigated loss to society?

Stakeholder participation: In addition to participation, stakeholders should be empowered to monitor their own performance like, for example, the members of a well-functioning Water User Association currently do.

Research question. What institutional arrangements can be made to accommodate self- regulation?

Planning the restoration process: What is the current process in planning a restoration intervention?

Research question. What steps can be taken to ensure that restoration interventions improve both ecosystem and social/community resilience by integrating human and natural processes?

Research question. Can we predict how effectively a restoration intervention will restore a particular bundle of socio-ecological services? Or worded differently, can we design a restoration intervention to restore a specific, required, bundle of socio-ecological services?

The effect of climate variability on systems dynamic models: The SD model did not examine the effect of climate variability on restoration interventions.

xix

Research question. How can the influence of climate variability on restoration interventions be included in a system dynamic model?

The economics of restoration: The SD model and underlying economic evaluation of restoration costs and benefits had to work on existing market prices for natural assets.

Research question. How would the economics for restoration change under a full cost accounting approach to goods and services delivered by well-functioning ecosystems?

Research question. What is the likely time period wherein transition towards full-cost accounting will take place? How will this influence results obtained in this research?

Table of contents

EXECUTIVE SUMMARY III

PROJECT OVERVIEW ...... III BACKGROUND ...... III OBJECTIVES ...... VI APPROACH ...... VI RESULTS ...... IX CONTRIBUTION OF THE RESEARCH ...... XII CONCLUSIONS ...... XV TECHNICAL / OPERATIONAL RECOMMENDATIONS ...... XV FUTURE RESEARCH ...... XIX

TABLE OF CONTENTS XXI

LIST OF TABLES XXVII

LIST OF FIGURES XXX

ABBREVIATIONS/ACRONYMS USED XXXII

ACKNOWLEDGEMENTS XXXIV

CHAPTER 1 INTRODUCTION: HYPOTHESIS, SCOPE AND METHOD OF WORK 1

1.1 SETTING THE SCENE ...... 1 1.2 SCOPE OF WORK ...... 1 1.3 WORK PROTOCOL ...... 2 1.4 SITE SELECTION ...... 4 1.5 STRUCTURE OF THE REPORT ...... 6

CHAPTER 2 LITERATURE REVIEW 7

2.1 INTRODUCTION...... 7 2.2 PURPOSE AND DELIVERABLES ...... 8 2.3 RESEARCH METHODS ...... 9 2.4 RESULTS ...... 14 2.4.1 GENERAL STATISTICS ...... 14 2.4.2 COMPARATIVE ANALYSIS: RESTORATION ECOLOGY AND THE OTHER 12 JOURNALS ...... 18 2.4.3 RELATIVE STATE OF NATIONAL ECONOMIC DEVELOPMENT AND RESTORATION ...... 23 2.4.4 RESTORATION, PAYMENTS FOR ECOSYSTEMS, AGRICULTURE AND IMPACT ON POLICY...... 27

xxi

2.5 DISCUSSION ...... 34 2.6 CONCLUSION ...... 36

CHAPTER 3 RESEARCH METHODOLOGICAL ISSUES WITH RESPECT TO SYSTEMS DYNAMIC MODELLING 38

3.1 INTRODUCTION...... 38 3.2 GENERIC STEPS IN SD MODELLING ...... 38 3.2.1 DEFINING THE PROBLEM ...... 39 3.2.2 GENERATING ALTERNATIVES ...... 39 3.2.3 EVALUATING ALTERNATIVES ...... 40 3.3 APPLICATIONS IN ENVIRONMENTAL-ECONOMIC SYSTEMS ...... 41 3.4 STEPS IN THE MODELLING PROCESS ...... 44 3.4.1 COMPARISON BETWEEN DIFFERENT APPROACHES ...... 44 3.4.2 CONCEPTUAL MODEL FORMULATION ...... 46 3.4.3 DEVELOPMENT OF A DYNAMIC HYPOTHESIS ...... 46 3.4.4 QUANTITATIVE MODEL SPECIFICATION ...... 47 3.4.5 MODEL VALIDATION ...... 49 3.4.6 MODEL USE ...... 51 3.4.7 TESTS FOR BUILDING CONFIDENCE IN MODELS ...... 51 3.5 MODEL FEATURES ...... 53 3.5.1 MODELLING SOFTWARE UTILISED ...... 53 3.5.2 BASIC BUILDING BLOCKS ...... 53 3.5.3 DYNAMIC BEHAVIOUR ...... 54 3.6 CONCLUSION ...... 57

CHAPTER 4 OUTCOMES OF THE SYSTEMS DYNAMIC MODELLING AND PORTFOLIO MAPPING PROCESS 58

4.1 INTRODUCTION...... 58 4.2 LITERATURE REVIEW/BACKGROUND ...... 58 4.2.1 SYSTEM DYNAMICS AND RESTORATION ...... 58 4.2.2 PRODUCT AND PROCESS INNOVATION ...... 59 4.3 DATA AND MATERIALS ...... 60 4.3.1 STUDY DESCRIPTION ...... 60 4.3.2 DATA SOURCES ...... 61 4.4 CONCEPTUAL MODEL ...... 62 4.5 RESEARCH METHOD ...... 65 4.6 RESULTS ...... 66 4.6.1 VALUE OF THE PAYOFF VARIABLE ...... 66 4.6.2 SYSTEM DYNAMICS MODEL ...... 67 4.6.3 PORTFOLIO MAPPING ...... 69 4.7 DISCUSSION ...... 72 4.8 CONCLUSION ...... 73

xxii

CHAPTER 5 DISCUSSION OF THE RESULTS 74

5.1 MESSAGE 1: ECONOMIC EVALUATION TOOLS AND SYSTEMS DYNAMIC MODELS ARE USEFUL TO INFORM DECISION- MAKING ON RESTORATION ...... 75 5.1.1 INTRODUCTION ...... 75 5.1.2 ECONOMIC TOOLS NEED TO BE USED WITH A SENSITIVITY TO DYNAMIC REALITIES ...... 76 5.1.3 ECONOMIC TOOLS NEED TO BE USED AS IF RISK AND UNCERTAINTY MATTERS ...... 77 5.2 MESSAGE 2: RESTORATION POTENTIAL ASSESSMENT PROTOCOL ...... 78 5.2.1 INTRODUCTION ...... 78 5.2.2 ASPECTS AND COMPLEXITIES THAT NEED TO BE CONSIDERED ...... 78 5.2.3 ADAPTIVE MANAGEMENT AS AN APPROACH ...... 81 5.2.4 TOWARDS GUIDELINES ...... 81 5.3 MESSAGE 3: RESTORATION IS AN INVESTMENT IN NATURAL CAPITAL AND AN INSURANCE PREMIUM AGAINST RISK ...... 83 5.3.1 INTRODUCTION ...... 83 5.3.2 BACKGROUND: WATER BANKS ...... 85 5.3.3 ANALYSIS: RIPARIAN ZONES’ INSURANCE CAPABILITIES – AND RESTORATION AS AN ADAPTATION STRATEGY ...... 86 5.3.4 CONCLUSION ...... 91 5.4 MESSAGE 4: COMPLYING WITH THE ENVIRONMENTAL RESERVE: AT WHOSE AND AT WHAT COST? ...... 92 5.5 MESSAGE 5: HOW TO DEAL WITH UNMITIGATED DAMAGE AND ITS COSTS TO SOCIETY ...... 97 5.5.1 INTRODUCTION ...... 97 5.5.2 TARGET SETTING (RESTORATION OBJECTIVES) IN ENVIRONMENTAL MANAGEMENT PLANNING ...... 98 5.5.3 UNMITIGATED DAMAGE ...... 99 5.5.4 MANAGING THE UNMITIGATED DAMAGE TO SOCIETY AS A RESULT OF TRANSFORMATION OF NATURAL LANDSCAPES ...... 99 5.6 MESSAGE 6: STAKEHOLDER PARTICIPATION: AN ATTEMPT TO CLOSE THE GAP BETWEEN RESEARCH AND ACTION ...... 102 5.6.1 INTRODUCTION & BACKGROUND ...... 102 5.6.2 ANALYSIS: THE OUTCOMES ...... 105 5.6.3 CONCLUSION: THE WAY FORWARD ...... 106 5.7 MESSAGE 7: POLICY RECOMMENDATIONS TOWARDS THE ACCELERATION OF THE RESTORATION OF NATURAL CAPITAL ...... 107 5.7.1 INTRODUCTION ...... 107 5.7.2 GENERAL POLICY RECOMMENDATIONS AND SOME OBSERVATIONS ...... 110 5.7.3 SPECIFIC POLICY RECOMMENDATION ...... 115 5.7.4 CONCLUSION ...... 120

CHAPTER 6 CONCLUSIONS, RECOMMENDATIONS AND FUTURE RESEARCH 121

6.1 INTRODUCTION...... 121 6.2 MAIN CONCLUSIONS ...... 122 6.3 RECOMMENDATIONS ...... 124 TECHNICAL / OPERATIONAL RECOMMENDATIONS ...... 124 FUTURE RESEARCH ...... 128

xxiii

REFERENCES 130

THE FOLLOWING ANNEXURES APPEAR ON THE ENCLOSED CD

ANNEXURE A PARAMETERS USED IN SYSTEMS DYNAMIC MODEL 141

1 AGULHAS SITE ...... 141 A PARAMETERS ...... 141 B EQUATIONS ...... 144 2 BEAUFORT WEST SITE ...... 148 A PARAMETERS ...... 148 B EQUATIONS ...... 149 3 DRAKENSBERG SITE ...... 151 A PARAMETERS ...... 151 B EQUATIONS ...... 152 4 NAMAQUALAND SITE ...... 154 A PARAMETERS ...... 154 B EQUATIONS ...... 156 5 KROMME SITE ...... 159 A PARAMETERS ...... 159 B EQUATIONS ...... 161 6 LEPHALALE SITE ...... 163 A PARAMETERS ...... 163 B EQUATIONS ...... 164 7 OUDTSHOORN SITE ...... 167 A PARAMETERS ...... 167 B EQUATIONS ...... 169 8 SAND RIVER SITE ...... 171 A PARAMETERS ...... 171 B EQUATIONS ...... 172

ANNEXURE B NAMAKWA SANDS 175

B.1 MONITORING ECOLOGICAL REHABILITATION ON A COASTAL MINERAL SANDS MINE IN NAMAQUALAND, SOUTH AFRICA ...... 175 B.2 A FINANCIAL AND COST BENEFIT ANALYSIS OF VELD REHABILITATION AFTER SAND DUNE MINING ...... 177

ANNEXURE C OUDTSHOORN 191

C.1 THE EFFECT OF REHABILITATION ON ECOSYSTEM SERVICES IN THE SEMI-ARID SUCCULENT KAROO LOWLANDS OF THE LITTLE KAROO, SOUTH AFRICA ...... 191

xxiv

C.2 A FINANCIAL COST-BENEFIT ANALYSIS OF THE IMPLEMENTATION OF A SMALL-CAMP SYSTEM IN OSTRICH FARMING TO ALLOW VELD RESTORATION ...... 193

ANNEXURE D BEAUFORT WEST 194

D.1 IMPACT OF PROSOPIS (MESQUITE) INVASION AND CLEARING ON ECOSYSTEM STRUCTURE, FUNCTION AND AGRICULTURAL PRODUCTIVITY IN SEMI-ARID NAMA KAROO RANGELAND, SOUTH AFRICA ...... 194 D.2 COST-BENEFIT ANALYSES OF ALIEN REMOVAL AND NATURAL CAPITAL RESTORATION: CASE STUDIES OF THE AGULHAS PLAIN AND BEAUFORT WEST LOCAL MUNICIPALITY ...... 195

ANNEXURE E LEPHALALE 242

E.1 RESTORATION OF AN ENCROACHED SEMI-ARID SOUTHERN AFRICAN SAVANNA. AN EVALUATION OF DIFFERENT TREE THINNING TREATMENTS ...... 242 E.2 ECONOMICS OF RESTORATION: THE ECONOMIC EVALUATION OF RESTORATION ON THE ECOSYSTEM SERVICES OF ENCROACHED SAVANNA IN LEPHALALE ...... 243

ANNEXURE F AGULHAS 276

F.1 DETERMINING THE HYDROLOGICAL BENEFITS OF CLEARING INVASIVE ALIEN VEGETATION ON THE AGULHAS PLAIN, SOUTH AFRICA ...... 276 F.2 COST-BENEFIT ANALYSES OF ALIEN REMOVAL AND NATURAL CAPITAL RESTORATION: CASE STUDIES OF THE AGULHAS PLAIN AND BEAUFORT WEST LOCAL MUNICIPALITY ...... 277

ANNEXURE G KROMME 278

G.1 AN ECOLOGICAL AND HYDROLOGICAL EVALUATION OF THE EFFECTS OF RESTORATION ON ECOSYSTEM SERVICES IN THE KROMME RIVER SYSTEM, SOUTH AFRICA ...... 278 G.2 WATER SUPPLY IN THE EASTERN CAPE: AN ECONOMIC CASE STUDY OF LAND REHABILITATION IN THE KROMME RIVER CATCHMENT ...... 279

ANNEXURE H DRAKENSBERG 281

H.1 AN ASSESSMENT OF THE ECOLOGICAL IMPACTS OF COMMUNITY-BASED REHABILITATION ON COMMUNAL GRASSLANDS IN THE DRAKENSBERG FOOTHILLS ...... 281 H.2 A COST BENEFIT ANALYSIS OF THE IMPACTS OF GRASSLAND REHABILITATION IN OKHOMBE, DRAKENSBERG .... 282

ANNEXURE I SAND RIVER 293

I.1 A COST-BENEFIT ANALYSIS OF THE IMPACTS OF WATER SUPPLY MANAGEMENT IN THE SAND RIVER CATCHMENT 293

ANNEXURE J MODELLING THE ECOLOGICAL-ECONOMIC IMPACTS OF RESTORING NATURAL CAPITAL, WITH A SPECIAL FOCUS ON WATER AND AGRICULTURE, AT EIGHT SITES IN SOUTH AFRICA 307 xxv

ANNEXURE K ASSET RESEARCH MODEL OF LEARNING AND CAPACITY-BUILDING 309

INTRODUCTION ...... 309 ASSET RESEARCH DEFINED ...... 309 ASSET RESEARCH OPERATIONAL MODEL ...... 310

xxvi

List of Tables

Table a Description of restoration study/project sites ...... v Table b List of students involved in the study by site and discipline ...... ix Table c Summary of present values for the different restoration sites (Rand million) ...... x Table d Summary of projects classified by type ...... xii Table 1.1 List of students involved in the study by site and discipline ...... 6 Table 2.1 Variables and its categories used in literature review ...... 10 Table 2.2 List of journals scrutinised ...... 12 Table 2.3 Number of papers scrutinised and the number dealing with either restoration or rehabilitation per year ...... 15 Table 2.4 Restoration or rehabilitation papers by ecosystem type ...... 16 Table 2.5 Number of papers by restoration approach ...... 17 Table 2.6 Number of papers by restoration approach and the use of instrumentation ...... 17 Table 2.7 Comparison between Restoration Ecology and 12 other journals ...... 20 Table 2.8 Comparison between Restoration Ecology and 12 other journals concerning the number of papers by scale1 ...... 22 Table 2.9 Comparison between Restoration Ecology and 12 other journals concerning PES and scale ...... 22 Table 2.10 Distribution of studies by ecosystem function in terms of the economic development status of the host country ...... 24 Table 2.11 Distribution of studies by contribution to human well-being in terms of the economic development status of the host country ...... 25 Table 2.12 Distribution of studies by policy outcome in terms of the economic development status of the host country ...... 25 Table 2.13 Distribution of studies with an agriculture link in terms of the economic development status of the host country ...... 26 Table 2.14 Distribution of studies with a PES link in terms of the economic development status of the host country ...... 27 Table 2.15 Agricultural link with restoration projects in varying types of ecosystems ...... 28 Table 2.16 The development of markets for ecosystem services for restoration-related activities ...... 29 Table 2.17 Markets for ecosystem services by ecosystem type ...... 29 Table 2.18 Comparison between PES and agriculture ...... 30 Table 2.19 Treatment of agriculture related to restoration in various journals ...... 31 Table 2.20 Different treatments of agriculture between Restoration Ecology and the other 12 journals ...... 32 Table 2.21 Monitoring tools used during ecosystem restoration ...... 32 Table 3.1 Stages in system problem solving ...... 39 Table 3.2 Comparisons between methods of problem solving ...... 39 Table 3.3 Sustainability problems and relevance for system dynamics modelling ...... 40 Table 3.4 System dynamics case studies referenced for geographical map, sorted by year ...... 43 Table 3.5 Comparison between Systems Engineering (SE) and System Dynamics (SD) methodologies ...... 44 Table 3.6 Steps in the model building process ...... 45

xxvii

Table 3.7 Summary of validation tests conducted by different authors ...... 50 Table 3.8 Basic entities in a stock flow diagram ...... 53 Table 3.9 Three types of variables used in the model ...... 54 Table 4.1 Landuse and ecosystem characteristics of the eight study sites ...... 61 Table 4.2 Monte Carlo summary statistics for output variable (NPV, t=50, 2060) ...... 65 Table 4.3 Summary of PVs for the different restoration sites (R million) ...... 66 Table 4.4 Summary of projects classified by type ...... 72 Table 5.1 List or riparian ecosystems services ...... 87 Table 5.2 Impact of natural resource management regime changes on net present value and the water flow change in the Sand River Catchment. All acronyms are defined below the table...... 95 Table 5.3 Summary of projects classified by type ...... 108 Table 5.4 Main findings and recommendations of the eight restoration sites ...... 116 *Species of seed sown – Morestȇr Farm, Little Karoo...... 119 Table 6.1 Summary of projects classified by type of risk and the likelihood of returns at each of the eight study sites from a market for ecosystem goods and services’ perspective* ...... 123 Table B1 The cost per ha of veld rehabilitation for the year 2009 ...... 182 Table B2 The profitability of sheep production on unmined land and rehabilitated land ...... 187 Table B3 Ecosystem services present at Namakwa Sands ...... 188 Table D1 Private and social benefits and costs ...... 208 Table D2 Discounted total costs and benefits (R millions) ...... 213 Table D3 Required value per cubic metre of water ...... 213 Table D4 Average water values per water use in the Breede WMA ...... 213 Table D5 Water valuation studies (Adapted from Moolman, Blignaut & van Eyden, 2006) ...... 214 Table D6 REFIT 2009 Phase 2 (NERSA, 2009) ...... 223 Table D7 Required profit margin across 20 years ...... 226 Table D8 Required profit margin across 20 years; including cost recovery ...... 227 Table D9 Discounted costs and benefits over 20 years ...... 228 Table E1 REFIT Tariffs R/kWh (NERSA 2009) ...... 247 Table E2 Expected ecosystems services of the restored savanna landscape, adapted by the Millennium Ecosystem Assessment (2003) ...... 249 Table E3 Assumptions pertaining to details of farm types ...... 251 Table E4 Plot densities and corresponding ETTEs ...... 252 Table E5 Current density options for Lephalale ...... 252 Table E6 Associated dry matter per hectare ...... 252 Table E7 Change in ETTE's and Grazing Capacities ...... 253 Table E8 Regression results ...... 253 Table E9 Change in grazing capacity as a result of restoration ...... 255 Table E10 Changes in ETTE’s and Browsing Capacity ...... 255 Table E11 Regression Results – Browsing Capacity ...... 256 Table E12 Change in browsing capacity as a result of restoration ...... 257 Table E13 A comparison of Musina to the study site, Durban...... 259 Table E14 Stocking rates for Musina Experimental farm 1992-1998 (Bothma & du Toit, 2010) ...... 259 Table E15 Gross Margin per hectare ...... 259 Table E16 Gross Margin per LSU ...... 260 xxviii

Table E17 Cattle Farms – Change in Gross Margin (R/ha) ...... 260 Table E18 Farm and Municipality benefits – cattle ...... 261 Table E19 Economic Value of optimal density level (4000ETTE/ha) ...... 261 Table E20 The gross margins of wildlife at different density levels ...... 261 Table E21 Net economic changes for wildlife per hectare ...... 262 Table E22 Farm and municipal benefits – game ...... 262 Table E23 Average Lephalale farm – economic benefits of restoration ...... 262 Table E24 Clearing assumptions ...... 263 Table E25 Costs associated with each scenario ...... 265 Table E26 Aggregated costs of clearing ...... 265 Table E27 Relationship between removed ETTE and dry matter ...... 266 Table E28 Beef Budget ...... 271 Table E29 CBA: "LOW" Density ...... 272 Table E30 CBA: "MEDIUM" Density ...... 272 Table E31 CBA: "HIGH" Density ...... 272 Table E32 CBA: "VERY HIGH" Density ...... 272 Table E33 Cost Recovery: Renewable Energy Farm level ...... 273 Table E34 CBA results with cost recovery cost option ...... 275 Table E35 CBA and Cost recovery option: fuel wood ...... 275 Table H1 Financial cost benefit analysis for restoration ...... 289 Table H2 NPV, IRR and LNPV for soil stability and financial NPV ...... 290 Table H3 NPV, IRR and LNPV for all ecosystem benefits ...... 290 Table H4 Contribution to total ecosystem benefits ...... 291 Table H5 NPV of ecosystem benefits net of costs in Okhombe ...... 291 Table I1 Land use within the Sand river catchment: ...... 294 Table I2 Water Balance estimates for the Sand River catchment ...... 296 Table I3 Water demand estimates (m3/ha/a) for alternative landuses ...... 297 Table I4 Current and proposed water allocation ...... 302

xxix

List of Figures

Figure a Geographical distribution of case studies ...... v Figure b Cumulative NPV (Rand/hectare) for each of the different study sites ...... x Figure c Portfolio map for different ecosystem services (bubble size indicates resources committed to it). Study sites: Ag=Agulhas; BW=Beaufort West; D=Drakensberg; Ka=Kromme (agriculture); Kna=Kromme (no agric); Lp=Lephalale; N=Namaqualand; S=Sand ...... xi Figure d When to use markets to assist in restoration activities ...... xiv Figure 1.1 Geographic distribution of the study sites ...... 5 Figure 2.1 The restoration method utilised, ecosystem function and the constituents of well-being being addressed or affected ...... 18 Figure 2.2 Institutional demarcation of those conducting restoration projects as published in Restoration Ecology ...... 21 Figure 2.3 Distribution of studies by the economic status of the host country ...... 23 Figure 2.4 The indicated policy outcome (a) and intensity (b) of restoration-related studies ...... 33 Figure 3.1 A conceptual model of landuse change in Western Australia ...... 42 Figure 3.2 Iterative model building process ...... 45 Figure 3.3 Simple causal loop diagram to illustrated balancing and reinforcing loops ...... 48 Figure 3.4 Tests appropriate at different stages of the model building process ...... 52 Figure 3.5 A System Dynamics model for fish population dynamics ...... 54 Figure 3.6 Examples of dynamic behaviour in systems models ...... 56 Figure 4.1 Geographical distribution of case studies ...... 60 Figure 4.2 Conceptual model ...... 62 Figure 4.3 Casual loop diagram ...... 64 Figure 4.4 Cumulative NPV (R/hectare) for each of the different study sites ...... 67 Figure 4.5 Illustration of the dynamics of the systems model ...... 67 Figure 4.6 Monte Carlo simulations for: a. changes in profitability of biomass electricity plant b. changes in the water value ...... 68 Figure 4.7 Monte Carlo simulations for price sensitivity of water for a. Drakensberg b. Beaufort West ...... 69 Figure 4.8 Portfolio map for different ecosystem services: bubble size indicates resources committed to it (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn) ...... 70 Figure 4.9 Portfolio map for different ecosystem services: bubble size indicates standard deviations (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn) ...... 71 Figure 4.10 Portfolio map for different ecosystem services: bubble size indicates coefficient of variation ...... 72 Figure 5.1 The three dimensions and underlying factors that need to be considered when evaluating the potential for a programme or project aimed at restoring natural capital...... 80

xxx

Figure 5.2 Integrated portfolio map for different ecosystem services (from top to bottom, bubble sizes indicating: Resources committed, standard deviation, coefficient of variation) ...... 109 Figure B1 Location of the mining activities and rehabilitation programme at Brand-se-Baai, South Africa ...... 179 Figure B2 Extent of land under rehabilitation over time at Brand-se-Baai in Namaqua ...... 181 Figure B3 Progressive rehabilitation of the mined land ...... 181 Figure B4 Basic structure of a farm model ...... 186 Figure D1 Cape Floristic Region, South Africa ...... 203 Figure D2 The Agulhas Plain (Nowell, 2010) ...... 203 Figure D3 Vegetation types of the Agulhas Plain (Nowell, 2010) ...... 204 Figure D4 Location of the study site (Municipal Demarcation Board, 2006) ...... 205 Figure D5 Breede Water Management Area ...... 207 Figure D6 Schematic representation of EGS affected by alien removal and restoration ...... 207 Figure D7 Total private income from harvesting ...... 211 Figure D8 Clearing and restoration cost for the entire Plain ...... 211 Figure D9 Private NPV per annum over 20 years ...... 212 Figure D10 Central Lower Nama Karoo within the Nama Karoo ...... 215 Figure D11 Central Karoo District Municipality ...... 216 Figure D12 Prosopis invasion in Beaufort West Local Municipality (Source: Wannenburgh, 2010) ...... 218 Figure D13 Water supply to Beaufort West town (Beaufort West Local Municipality, 2010)...... 218 Figure D14 Theoretical illustration of water valuation ...... 221 Figure D15 Tariff structure in Beaufort West Local Municipality ...... 223 Figure D16 NPV for cost recovery project lifetimes of up to twenty years (discount rates in parentheses) ...... 226 Figure E1 Limpopo Biome Map illustrating the savanna ecosystem in the Waterberg (DEAT 2010) ...... 246 Figure E2 Sectoral analysis of employment for Lephalale (IDP 2009/10) ...... 246 Figure E2 Relationship between grazing capacity and bush encroachmen ...... 254 Figure E3 Relationship between browsing capacity and bush density ...... 256 Figure E4 Relationship between grazing capacity and % encroachment ...... 257 Figure E5 Relationship between browsing capacity and % encroachment ...... 258 Figure E6 Cost Recovery "Low" Current Density ...... 266 Figure E7 Cost Recovery and Cost of Clearing...... 267 Figure H1 Location of Okhombe within KwaZulu-Natal Province and South Africa ...... 283 Figure H2 Soil yield per millimetre of rainfall across six sample years ...... 286 Figure H3 Sediment yield per millimetre rainfall for a) Rehabilitated area and b) eroded area ...... 287 Figure I1 Location of the SRC within the Sabie-Sand catchment, South Africa ...... 294 Figure I2 Additional water yields and levelised NPV under the MWD scenario ...... 303

xxxi

Abbreviations/acronyms used

AWD Allocatable Water Demand BCR Benefit Cost Ratio BTP Bushbuckridge Transfer Pipeline CARA Conservation and Agricultural Resources Act CLD Causal Loop Diagram CV Coefficient Of Variation DWA Department Of Water Affairs EGS Ecosystem Goods And Services EMP Environmental Management Plans ESA Ecological Society of America ET Evapotranspiration ETTE Evapotranspiration Tree Equivalent GEM General Equilibrium Modelling IAC Irrigated Annual Crops IAP Invasive Alien Plant IAPS Interested And Affected Parties IPC Irrigated Permanent Crops IRR Internal Rate Of Return LSU Large stock unit MAP Mean Annual Precipitation MWD Maximum Water Demand NC Neoclassical NEMA National Environmental Management Act NPV Net Present Value NWA National Water Act PES Payments For Ecosystem Services PLE Personal Learning Edition PM Portfolio Mapping PV Present Value RA Risk Analysis Regional Economic System Dynamics Model For The Restoration Of RESTORE-P Ecosystems And Project Prioritisation RNC Restoring Natural Capital RoD Record of Decision SAAB South African Association Of Botanists SD Systems Dynamics SE Systems Engineering SER Society For Ecological Restoration SSGR Sabie Sand Game Reserve SSU Small stock unit SU University of Stellenbosch UCT University of Cape Town UFS University of the Free State

xxxii

UKZN University of KwaZulu-Natal WA Western Australia WRC Water Research Commission

xxxiii

Acknowledgements

ASSET Research would like to thank the Water Research Commission (primarily) and the Natural Resource Management programmes (secondary) for the financial support provided. We wish to thank also the following reference group members for their input during the project:

Dr AJ Sanewe Water Research Commission (Chairman) Dr GR Backeberg Water Research Commission Dr TM Everson University of KwaZulu-Natal Dr PJ Dye University of Witwatersrand Prof GPW Jewitt University of KwaZulu-Natal Dr C Marais Department of Environmental Affairs Dr J Turpie University of Cape Town Dr H van Zyl Independent Economic Consultants

ASSET Research also gratefully acknowledges the following institutions for additional funding and/or assistance with the project:

• Exxaro • WfW • Ostrich Business Chamber • AWARD • Flower Valley Conservation Trust • Elsenburg • CIB, University of Stellenbosch • University of KwaZulu-Natal

ASSET Research also gratefully acknowledges contributions by the following supervisors:

• Prof Nico Smit • Prof Theo Kleynhans • Prof Geoff Antrobus • Prof Tony Leiman • Dr Terry Everson

xxxiv

CHAPTER 1 INTRODUCTION: HYPOTHESIS, SCOPE AND METHOD OF WORK

James Blignaut, Martin de Wit, Sue Milton, Karen Esler, and David le Maitre

1.1 Setting the scene

South Africa has a proud history of restoring natural capital (RNC) but analyses of restoration have often been neglected. A meta-analysis of the ecological, hydrological, and economic impacts of restoration across a range of contrasting sites and contexts is lacking. This study aimed to address this deficiency and determine the tangible contributions of restoration. We focus on existing restoration sites and monitor and evaluate the ecological, hydrological and socio-economic impacts restoration might have had at those different sites. We use a carefully selected set of ecological, hydrological and socio-economic parameters to test the following hypothesis at different sites:

The impact of RNC is, firstly, on the flow of a suite of ecosystem goods and services and, secondly, on the economy through enhanced income, job creation (during and after restoration) and through improved returns from the land. In effect, RNC is no different from the capital expenditure on any other project, whether in construction, engineering and agriculture. The return to the land and the value of the environmental services of the resulting flows is the annual stream of benefits delivered at an annual maintenance and operation cost. In addition to the broad goal of improving the links between applied ecology and sustainable economic development, this study is of value to the rural, often agricultural, as well as urban and recreational users of ecosystem goods and services. This study assessed the impact of RNC on these aspects by considering eight existing case studies. This study also defines the parameters to use in measuring the success of future restoration projects, which will, by definition have an impact on the economy, health and the environment.

The study was conducted by ASSET Research (www.assetresearch.org.za). ASSET is a not-for-profit public benefit organisation that specialises in R&D and capacity-building in the field of environmental and ecological economics. Its primary objective is to offer students the opportunity to work in multidisciplinary teams on projects and, in the process, innovate, learn and contribute to the ecological, hydrological and economic knowledge base of the country. The study is therefore mainly conducted by students. In total ten masters theses and one doctoral thesis were produced covering three disciplines and involving three tertiary institutions; see Annexure J for a discussion on the ASSET research and capacity development model.

1.2 Scope of work

The focus of this study is on conducting a meta-analysis of the hydrological, ecological and socio- economic impacts of restoration at 8 different sites and no primary restoration implementation was 1

done. This implies evaluating and monitoring the impact “with” and “without” restoration at the 8 sites – see both the criteria for selecting the sites and the selected sites themselves below. For each restoration site there was also a control site, which in this case implies an un-restored or degraded site. The outcome of conducting the evaluation at such a large number of sites is the compilation of an integrated ecological-economic systems model. This will assist policy-makers to identify the best sites for future restoration and to model a range of scenarios under various assumptions, leading to the identification of the potential risks. From an operational perspective, the students worked in site- specific teams conducting their fieldwork with respect to hydrology, ecology and economics. Each student, whether in hydrology, ecology or economics, was responsible for either one or two sites with the objective of quantifying the ecosystem services and values (with and without restoration) to demonstrate the difference RNC makes.

1.3 Work protocol

Through the Working for Water, Woodlands, Wetlands and Fire projects – and various other restoration projects such as the mandatory restoration and rehabilitation of road servitudes and mine-dumps – South Africa has established itself as a country which has started to invest in the RNC. This is especially true among developing nations where South Africa is being used as an example and leader. However, despite this rich history in restoration, no meta-analysis has been done to assess restoration’s ecological, hydrological, and economic impacts across a range of contrasting sites and contexts. This study aimed to rectify this obvious deficiency.

The specific objectives of this study were to conduct ecological, hydrological, and socio-economic assessments to determine the impact of restoration at eight ecologically and socio-economically different restoration sites in comparison to degraded or un-restored areas in close proximity to the eight selected sites. Following this an integrated systems dynamic model on the likely impact of restoration on the ecology, hydrology and economy of notably agriculture, was developed.

Restoration impacts positively on the flow of a suite of ecosystem goods and services and on the economy. While some evidence exists on the ecological and hydrological implications of restoration for individual projects, their links to the economy and spanning various sites and biomes has not been done. There is also no clear understanding of how the benefits before and after restoration might affect agriculture through improved returns from the land. These benefits generally are believed to be very real and significant but not well understood. This study endeavoured to provide these links. In effect, restoration is no different from the capital expenditure on any project and the return to the land. The value of the environmental services of the ensuing flows (as a result of the capital expenditure) is the annual stream of benefits delivered at an annual maintenance and operation cost.

To achieve the stated objectives, this study selected an assemblage of ecological and socio- economic parameters to test the hypothesis stated above. The overall study and the studies at each of the eight sites were based on the generic design described below. The research team identified eight existing restoration sites that were well established and that had significant/sufficient supporting data and allocated the study sites to student teams. In so doing each student in restoration ecology carried out an ecological assessment of the impact of restoration on ecosystem function in his/her site. Together they covered all eight sites, asking the same questions and doing similar types of work. The

same applied to the hydrological and economic assessments. The basic field study design included a comparison of 4 types of (sufficiently similar) sites, namely:

In the sites where hydrological assessments were conducted the hydrological studies focused on how invasions and clearing have affected the hydrological functioning of a site, focusing particularly on:

• One study derived a basic catchment water balance based on the difference between the rainfall and the estimated transpiration and interception losses for sub-catchment units based on data on the vegetation structure and evaporation. • Then other studies assessed the relative quantities of water that become overland flow versus that infiltrating the soil using combinations of runoff plots, infiltrometers and measures of the soils porosity and proneness to form mineral crusts (e.g. Mills & Fey 2004a, b). • The site-based studies collected a number of indices of site hydrological function based primarily on the approaches developed by David Tongway and others; these use various indices to quantify the nature of the soil surface, vegetation (e.g. basal cover) and indicators of surface water flow and sediment transport. These indices were developed largely from studies of patchy semi-arid vegetation and reflect the basic hypothesis that “pristine” landscapes retain resources and degraded ones leak resources (Ludwig et al. 1997 and subsequent papers; Belnap et al. 2008). These basic hypotheses and the indices were tested for local applications incorporating local experience and knowledge (see Le Maitre et al. 2007 on the Little Karoo).

The methods were standardised across the different studies as far as possible so that all the hydrological function indices and vegetation measurements were collected on all the sites. The actual measurements of infiltration-related variables varied from site to site depending on the conditions, but a minimum set was collected by all the site-based studies. The study was run in close conjunction with the ecological study and contributed to, and used, the historical and ecological information collected for each site. The students involved in the hydrological studies were required to review the available methods and to get expert opinions on their applicability before the final set of measurements was chosen. The hydrological studies were aimed at testing a range of relatively straightforward indices of site water balance and hydrological function that can be used by land managers in a range of vegetation types and conditions. There are soil erosion models that could have been used but there are still many issues to be addressed (Boardman 2006) so erosion modelling was not envisaged.

There were two components to the site-based ecological assessments, namely:

• a field component focusing on understanding the impact that the restoration activity has had on ecosystem diversity, structure and function.

Each ecology student used a combination of line transects and plots to assess the following variables where applicable:

• cover of alien and indigenous plant species • species richness of alien and indigenous plant species • vegetation structure • resilience (changes in cover between seasons) • grazing capacity (ha/LSU) • soil structure • assess the change in ecosystem function before and after restoration

From the above, the impact restoration has had on the flow and generation of ecosystem goods and services was determined, with a special reference to its value for agriculture.

The team of economic students worked on the same eight sites and used the information generated by the ecological and hydrological studies, in combination with local level surveys/questionnaires and/or group sessions, to determine the socio-economic value of RNC in each case. This was done to determine the economic value of the site before and after restoration, or in comparison with a control site/area. The socio-economic assessments were done by means of surveying the local interested and affected parties. This implies that in most cases both an economic and financial cost-benefit analysis was done to determine the economic viability and feasibility of restoration. This portion of the study explicitly looked at the possibility and/or feasibility of embarking on a trade for ecosystem services and the benefits of restoration to agriculture.

1.4 Site selection

Given the range of sites where restoration is or has been conducted in South Africa, it was necessary to establish criteria to determine which eight study sites should be selected within the context of this study. The following list of important considerations was decided upon:

1. Student safety 2. Equipment security 3. Sites representing different biomes 4. The presence of existing and collaborating networks/individuals 5. Availability of historic data on the site 6. The ease of replicating the study – the more generic the site/area the better 7. Accessibility of the site 8. Suitable reference sites in each case 9. Ability to use a consistent/comparable research method across the study sites 10. Site has to have strong linkages to agriculture production systems

11. Social upliftment potential 12. Economic development potential 13. Possibility to institute a market for ecosystem goods and services 14. Location/distance between sites

Guided by this list and the outcomes of a process of review and stakeholder consultation, we selected the following eight sites (see also Figure 1.1 for a perspective on the geographic distribution):

1. Restoration of sand dunes in the succulent karoo following open-pit surface mining in the Namaqualand; 2. Restoration following overgrazing by ostriches in the succulent karoo near Oudtshoorn; 3. Clearing of invasive alien plants (notably Prosopis) in the Nama karoo near Beaufort-West; 4. Bush-thinning (and combating bush encroachment) in the bushveld/savanna near Lephalale; 5. Clearing of invasive alien plants in fynbos ecosystems on the Agulhas plains; 6. Clearing of invasive alien plants in the fynbos and riparian ecosystems of the Kromme river system; 7. Restoration of a communal grassland system following overgrazing near the Okhombe village in the Drakensberg; and 8. Removal of exotic plantation forestry within the Sand River catchment.

Figure 1.1 Geographic distribution of the study sites Note: Sites are ordered according to the average rainfall from low to high.

Table 1.1 List of students involved in the study by site and discipline

Registration Qualification Student University Site Discipline Year (degree) Agulhas and Helanya Vlok 2009 MComm Economics SU Economics Oudtshoorn MSc Conservation Ecology with Marco Pauw 2009 SU Namaqualand Ecology hydrological aspects MSc Conservation Ecology with Megan Nowell 2009 SU Agulhas Ecology hydrological aspects Thabisisani MSc Conservation Ecology with 2009 SU Beaufort West Ndhlovu Ecology hydrological aspects MSc Conservation Ecology with Petra de Abreu 2009 UCT Oudtshoorn Biology hydrological aspects MSc Agricultural Oudtshoorn and Worship Mugido 2009 SU Economics Economics Namaqualand MSc Conservation Ecology with Alanna Rebelo 2010 SU Kromme river Ecology hydrological aspects Kromme river Katie Gull 2010 MSc Economics UCT Economics and Lephalale Ecology with Jacques Cloete 2010 MSc Natuurlewe UFS Lephalale hydrological aspects MSc Conservation Ecology with Dane Marx 2010 UCT Drakensberg Biology hydrological aspects Drakensberg, Sand river and Douglas Crookes 2009 PhD SU Economics systems-dynamic modelling

1.5 Structure of the report

Chapter 2 captures the results of a literature review covering some 20,000 papers in the academic literature with respect to restoration. This provided the baseline information for the remainder of the study. The review of both local and international literature assessed the linkages between economic development and the RNC. Chapter 3 focuses on the research method with specific reference to system dynamics modelling. Chapter 4 contains the study results with Chapter 5 reflecting on some of the key messages and emerging issues based on the research. Chapter 6 concludes the report. The information in the report is supported by several Annexures.

CHAPTER 2 LITERATURE REVIEW1

James Blignaut, Martin de Wit, Sue Milton, Karen Esler, David le Maitre, Steve Mitchell, Douglas Crookes, Alanna Rebelo

2.1 Introduction

As discussed in the inception report, dated 27 June 2008, South Africa has a proud history of restoring natural capital (RNC) but no one has yet conducted a meta-analysis of the ecological, hydrological, and economic impacts and plausible benefits to society of RNC across a range of contrasting sites and contexts. The present study aimed to address this deficiency and to determine the tangible contributions, or outcomes, of RNC to human society, and “socio-ecological systems”, as perceived through the development of markets for ecosystem services.

Markets for ecosystem services are reasonable indicators for measuring the outcomes of ecological restoration and RNC for socio-ecological systems involving people and their systems, and ‘natural’ ecosystems upon which human societies and economies depend. This is because human societies and economic markets are simply social constructs built on the desires and wishes of people. Therefore, if there is strong evidence in the scientific literature that markets are being actively pursued and linked, or integrated, with the act of RNC, then it clearly signals the importance of the activity. Such evidence would support the notion that the outcomes of RNC are of sufficient value to people to justify the maintenance of such a social construct, i.e., an investment in RNC. Failure to develop such a market indicates either that there is insufficient value, or that scientists, policy-makers, and society in general, have failed to recognise the value of restoration for human livelihoods and well-being. There have been thousands of eco-restoration and RNC projects, and there is an extensive literature on the ecological theory behind them. However, despite a growing interest and scientific literature, it is clear that the direct value to society of the practice and outcomes of restoration or RNC have yet to be properly studied or described. A lack of markets and of discussion of market potential could be seen as a signal of a failure to communicate the benefits and effectively involve society and institutions in restoration initiatives. Here we provide a preliminary discussion of a study that tests this idea.

In addition to the immediate objective of the study as described above, the larger, 5-year long, study RNC endeavoured to consider outcomes of restoration activities across a range of sites and to monitor and evaluate the ecological, hydrological and socio-economic impacts restoration may have

1 A selection of the findings of this research has been published in: Aronson, J., Blignaut, J., Milton, S., le Maitre, D., Esler, K., Limouzin, A., Fontaine, C., de Wit, M., Mugido, W., Prinsloo, P., van der Elst, L. and Lederer, N. 2010. Why restore? A Meta-analysis of Recent Papers (2000-2008) in Restoration Ecology and 12 other Scientific Journals. Restoration Ecology, 18(2):143-154.

The following manuscript containing research findings from this study has also been submitted for publication: Blignaut, J., Esler, K., de Wit, M., le Maitre, D., Milton, S. and Aronson, J. Economic development and the restoration of natural capital. Current Opinion in environmental Sustainability. (Under review)

had at those different sites. It used a carefully selected set of ecological, hydrological and socio- economic parameters to test the following hypothesis at different sites:

RNC improves water flow and water quality, land productivity, in some instances sequesters more carbon, and, in general, improves both the socio-economic value of the land in situ, and in the surroundings of the restoration site, as well as the agricultural potential of the land.

The outcomes of restoration are changes in the flow of a suite of ecosystem goods and services, changes in the local economy through job creation during and after restoration and, thirdly, through improved returns from the land. In effect, RNC is no different from the capital expenditure on any industrial or urban or rural development project. The return to the land, and the value of the environmental services of the resulting flows, is the annual stream of benefits delivered at an annual maintenance and operation cost. In addition to the broad goal of improving the links between applied ecology and sustainable economic development, this study will be of significant value to agriculture because this sector is the primary beneficiary of the improvements in human well-being and ecosystem function. This study will also help to define the parameters to use in measuring the success of future restoration projects which will, by definition, have an impact on the economy, safety and the environment. By and large, it is on the outcome of these parameters that society can develop markets, since markets desire a degree of rigour and discipline based on an inherent understanding of, and confidence in, the outcome of specific needs.

In this report we provide the final results of an extensive – probably the most extensive ever – review of the international academic literature on ecological restoration with special emphasis and focus on payments for increases in ecosystem services provided by eco-restoration. Thus we employ the terms holistic ecological restoration, and RNC, in addition to ecological restoration, as these terms are defined in the Glossary in Annexure A. It should be noted that although it makes good sense to distinguish between restoration and rehabilitation, we have had to use them interchangeably since this distinction is not consistent in the literature. Occasionally we use the term restoration to indicate restoration and rehabilitation, following the example of the widely-cited Primer on Ecological Restoration, of the Society for Ecological Restoration International (SER 2002). We also sometimes use the term restoration to indicate the broader notion of RNC, but these distinctions will be clarified in forthcoming publications based on this research.

In what follows, we first define the purpose of this review, then explain the research methods employed and provide a detailed section on results. Finally, we provide a discussion section and conclusions.

2.2 Purpose and deliverables

The purpose of this literature review was to develop an understanding of:

Monitoring and evaluation in restoration programmes (RNC, as indicated above) and its implication for the development of markets for ecosystem goods and services.

The deliverables of this portion of the research project included, in addition to the submitted draft report and this the final report, at least one but more likely three or four published papers in peer-

reviewed journals reporting on the current status of restoration, its monitoring and evaluation, with regards to restoration-related markets for ecosystem goods and services.

The results of this study also formed the core of a plenary presentation made at the 35th annual conference of the South African Association of Botanists (SAAB) in January 2009.

2.3 Research methods

We began by identifying those papers in the scientific and public domain that bear directly on this study. To be included in the literature survey, papers had to relate to eco-restoration and/or rehabilitation, identify what forms of monitoring and evaluation were used and, lastly, indicate whether such research was in any way linked to either a market or a payment system for ecosystem goods and services.

We began with a pilot study of 528 academic papers published in peer-reviewed journals from January 1, 2000 onwards, obtained from established researchers in the field of restoration ecology and ecological economics. We searched these papers for the words “restoration” or “rehabilitation” in their titles, abstracts or keywords. When a paper was found that contained either of the above- mentioned two words, the paper was classified as a “hit”; if not the paper was classified as a “no hit”. The papers classified as “hits” were subsequently analysed according to a pre-determined list of variables and categories which is likely to provide us with a link to measure, tangibly, restoration’s contribution to human-sociological systems (see Table 2.1).

From the sample of 528 papers initially scrutinised, 115 were identified as being a “hit”. After analysing the list of “hits”, we observed that the majority of the relevant papers were from 13 journals. We decided to focus on those journals (see Table 2.2) because a general search would be too extensive and this set included a broad cross section of highly rated ecological journals as well as a key journal from the field of economics. Five journals – Agriculture, Ecosystems and Environment, Biological Conservation, Conservation Biology, Ecological Economics and Journal of Applied Ecology – of the 13 journals retained, are included in the list of 14 journals cited in an earlier survey of “top” ecology journals selected by Lawler et al. (2006).

Table 2.1 Variables and its categories used in literature review

Category Keywords & Definitions Paper descriptors Author, Year of publication, Title of the article, Journal, Location of the keyword identification (in the Title, Abstract or Keywords and combinations of these) Ecosystem in which the Grassland, forest, wood & savanna, shrubland, arid & semi-desert & desert, study was conducted rivers, other wetlands, marine & costal, urban, human modified & transformed, other or unclassified Restoration approach Active = implies that something was added or removed (e.g. reseeding, fertilizer, irrigation, plants) Passive = area was left to recover by itself Not specified Restoration method used Re-seeding, planting, succession (assisted or directed), others, and not specified Purpose of restoration; Supporting = a service such as pollination or seed dispersal that makes it type of ecosystem goods possible to produce crops & services affected (as Regulating = a service that moderates environmental extremes or stabilises per the categories used ecosystem components, dynamics and functions – e.g. control of floods, in the Millennium erosion, dust storms Ecosystem Assessment Provisioning = direct values of goods that can be harvested e.g. firewood, 2005) craft materials, meat Cultural = benefits that people get from visiting wild places – scenery, traditional rituals, relaxation, scientific information Constituents of wellbeing Material = food, wood, fish and other things people harvest from ecosystems addressed or affected Health = health benefits of natural environments including water purification, removal of toxins from the air Security = ways in which natural vegetation or functioning ecosystems protect our atmosphere or prevent disasters such as floods Social relations = ways in which natural environments contribute to our cultural and social lives Wellbeing impact Describe in words how the restoration improves life quality for people description Link to agricultural Does the restoration link with agricultural systems or practices (including crop systems or practices production, forestry, ranching)? – Yes / No. If Yes, what and how? Monitoring tools used Yes / No. If Yes, describe in words how the restoration was monitored. Instrumental = measuring something such as vegetation cover or chemicals. Interviews = asking people how they benefited or whether they saw a change. Scale of influence and This refers to the level of ecological organisation addressed by the restoration interventions project – whether at: Landscape (spatial interactions) = covering many habitats or communities Ecosystem (trophic interactions) = that the restoration influences plants, herbivores & predators, Community (inter-specific interactions)= restoration affects many organisms, or Population (re-introductions) = restoration focused on a single species Policy outcome or This refers to the effect of the study of the restoration or the restoration itself (research) on policy: locally (one settlement), nationally (whole country) or global recommendation (whole world)

Policy intensity (scale of The number of people directly or indirectly affected by the policy or the impact) importance of the policy for the way in which towns, nations or the world is run: None, minor, major. Country Country where restoration took place PES Payment for Environmental Service, Yes, No. This describes ways in which restoring an environmental to provide better services can be rewarded (for example with tax reductions). If yes, is the market actual or perceived? Perceived = a possible method of reward has been described, Actual = the reward method is functioning and that farmers, miners, etc. are actually receiving some payment or other benefits for doing the restoration.

Table 2.2 List of journals scrutinised

Journal Publisher Scope of Journal Specific Focus relevant to this analysis Ecological Elsevier Trans-disciplinary, linking ecology & economics Sustainable agriculture & development; renewable Economics (International resource management & conservation; integrating natural society for resources & environmental services into income & wealth Ecological accounts Economics, ISEE) Environmental Cambridge Intersection of environmental, resource & Theoretical and applied (policy) aspects of sustainable Development University Press development economics development Economics Restoration Ecology Blackwell Focusing on ecological restoration as defined as Experimental, observational & theoretical studies on “the process of assisting the recovery of an terrestrial, marine and freshwater systems ecosystem that has been degraded, damaged, or destroyed” (SER 2002) Ecological Elsevier Bridge between ecologists & engineers (eco- Eco-technology; synthetic ecology; bioengineering; Engineering technology) involved in designing ecosystems sustainable agro-ecology; habitat reconstruction; for mutual benefit of humans & nature restoration ecology; ecosystem rehabilitation; stream & river restoration; wetland restoration & construction; reclamation ecology Agriculture, Elsevier Interface between agro-ecosystems (crops, Agricultural landscape ecology & processes; papers that Ecosystem and pastures, livestock) & environment (energy, air, advance understanding on how to make agro-ecosystems Environment water, land) more diverse and sustainable Water SA Water Resources All branches of water science, technology & Water resources development; surface hydrology; geo- Commission engineering hydrology; environmental pollution control; water quality & (WRC) (South treatment; agricultural water science Africa) Journal of Forest Elsevier Links forest ecology with forest management Application of biological, ecological & social knowledge to Ecology and the management & conservation of man-made & natural Management forests Journal of Arid Elsevier Multi-disciplinary / interdisciplinary, research on Physical, biological & anthropological aspects of arid, environments all aspects of arid environments & their past, semi-arid & desert environments, including land use, present & future use conservation, land degradation & rehabilitation, techniques for monitoring & management Journal of Applied Blackwell Application of ecological concepts, theories, Application & development of improved strategies for the Ecology models & methods to management of biological conservation of wildlife; wildlife & habitat management; resources in broadest sense sustainable management of natural resources (including aquatic); management of pests & weeds 12

Environmental Springer Conservation of natural resources, protection of Ecology, ecological economics, environmental Management habitats & control of hazards, spanning field of engineering applied ecology without regard to traditional disciplinary boundaries Conservation Blackwell (Society Contributions to the study & preservation of Biology for Conservation species & habitats Biology) Biological Elsevier (Society Biological, sociological, and economic Theoretical & empirical investigations into the Conservation for Conservation dimensions of conservation and natural consequences of human actions for the diversity, structure Biology) resource management & function of terrestrial, aquatic or marine ecosystems, including restoration ecology, resource economics. Frontiers in Ecology Ecological Society All aspects of ecology, the environment, and and Environment of America (ESA) related disciplines

The review was then expanded to focus on all the papers published between January 2000 and September 2008 in these 13 journals. Based on the outcome of the above-described selection, we now provide a summary of the results.

2.4 Results

2.4.1 General statistics

The results are based on 19,547 papers published in 13 journals (listed in Table 2.2) over almost eight years, from January 2000 to October 2008. Of these, a total number of 1,575 “hits” were obtained, i.e. the published scientific papers that contained either the word restoration or rehabilitation in their titles, abstracts and/or keywords. The “hits” therefore represents 8,1% of the total number of papers scrutinised (see Table 2.3). This includes the papers of Restoration Ecology and Ecological Engineering, two journals aimed mainly at the ecological restoration (sensu lato) audience. It is hard to understand why only one in 12,5 papers deals with restoration given the state of environmental degradation world-wide. If Restoration Ecology and Ecological Engineering are excluded, then the number of “hits” is reduced to just 4,6% of the total number of papers, a hit-rate of only one in 21 papers! One explanation for this could be that academic ecology to date has focussed largely on the theory, functioning and dynamics of ecosystems in their natural, or near natural, states or on set-aside for conservation areas, rather than modified and transformed systems where people live and make a living from the land (Palmer et al. 2004). Although restoring (reassembling) ecosystems provides a significant test of whether ecological theory works in practice (Diamond 1985; Bradshaw 1987) have tended to favour basic science as this is generally rated more highly than applied science in academic institutions and by most funding agencies.

There is, however, an encouraging increase in publications in this field as the number of restoration papers published in both 2007 and 2008 are twice those published in 2000. This is substantiated by the fact that in 2007 and 2008 15% & 14% respectively of the total number of “hits” appeared, while only 7% appeared in 2000. This corresponds with the trend, reported by Ormerod (2003), editor of the Journal of Applied Ecology, who noted a very marked increase in papers on restoration in that distinguished, highly academic journal over the last 40 years, and especially in the period 1992-2002.

Table 2.3 Number of papers scrutinised and the number dealing with either restoration or rehabilitation per year

Total Total on restoration % of papers p.a. % of restoration scrutinised or rehabilitation dealing with papers (“hits”) restoration 2000 1,629 111 6.8% 7.0% 2001 1,686 109 6.5% 6.9% 2002 1,740 133 7.6% 8.4% 2003 2,098 154 7.3% 9.8% 2004 2,141 182 8.5% 11.6% 2005 2,391 228 9.5% 14.5% 2006 2,689 205 7.6% 13.0% 2007 2,506 239 9.5% 15.2% 2008* 2,667 214 8.0% 13.6% Total 19,547 1,575 8.1% 100.0% * Note that only the publications appearing in the first nine months of this year were included.

Table 2.4 provides an analysis of the number of papers per year by ecosystem type. Please note that it is possible for the number of observations to exceed 1,575 since some of the studies addressed or incorporated more than one ecosystem type. There is a definite clustering of papers dealing with forests, woodlands and savannas, and a second one focussed on rivers and wetlands. (In Table 2.2, therefore, these two clusters have been grouped together.) This is clearly indicated by the number of times over the study period that restoration-related papers concerned with these two ecosystem clusters exceeded 20%. For forests, woodlands and savannas that was the case in every year since 2000, and for rivers and wetlands it was so for 8 of the 9 study years. Studies in urban and transformed landscapes reached the 20% mark of the total only once, in 2001, while none of the other ecosystems reached 20% even once. There is, for example, a nine-fold and four-fold difference respectively between the papers focussing on arid and semi-arid ecosystems and marine and coastal areas, as compared to the cluster of forests, woodlands and savannas. This is a finding in accord with that of Lawler et al. (2006) who state, after reviewing 628 papers: “We found geographic gaps in conservation research, with marine, tundra, and desert biomes being studied less than other systems”. The importance of, and rapidly increasing need for, available and clean water is also emphasising the significance of restoration in rivers and wetlands. This is likely to become even more important in future.

Two factors may account for this emphasis on forested ecosystems: Firstly, there is a high proportion of papers from the USA, where restoration of abandoned agricultural lands and old field succession studies, particularly in the eastern deciduous forests, has been a major theme of ecological research going back to the works of Clements (1916) and Leopold (1949). A quick assessment of a selection of the journals indicates that studies from the USA comprise 52% of the 514 in Restoration Ecology, 25% of 186 in Biological Conservation, 58% of 48 in Conservation Biology and 40% of 258 in Forest Ecology and Management. An interesting exception is Agriculture, Ecosystems and Environment where only 6% of the 34 ‘hits’ were from the USA. Secondly, many of the other studies are from European countries (see section 2.4.3), where forests feature prominently in mythology and cultural heritage both as a benefit and as a threat, going back at least as far as the Classical Greeks and Romans. A modern echo of this can be seen in, for example, Tolkien’s “Lord of

the Rings” which is rooted in mediaeval mythology. Forest ecology therefore features prominently in theory and practice in Europe and North and South America (Wiersum 1995). A scientific extension of this interest is the virtually exclusive focus on woody vegetation cover as an index of land degradation by the UN Food and Agricultural Organisation for several decades (e.g. FAO 2006), an approach that does not work well in environments where woody plant cover is low. There is also a measure of bias in the journals where 258 (17%) of the hits were from Forest Ecology & Management – where forest, woodland or shrubland ecosystems comprise 70% of the total identified – and for Restoration Ecology (514 hits or 33%) the corresponding value is 30%.

Table 2.4 Restoration or rehabilitation papers by ecosystem type

Grass- Forests, Shrub- Arid and Rivers Marine Urban Unclass Total lands woodlands lands semi-arid and and and -ified and savanna ecosystems wetlands coastal trans- areas formed 2000 11 (7) 40 (26) 2 (1) 5 (3) 59 (39) 10 (7) 20 (13) 5 (3) 152 (100) 2001 9 (7) 31 (24) 6 (5) 1 (1) 40 (31) 5 (4) 29 (22) 10 (8) 131 (100) 2002 20 (11) 42 (23) 9 (5) 6 (3) 65 (35) 18 (10) 23 (12) 3 (2) 186 (100) 2003 32 (17) 39 (21) 16(8) 6 (3) 39 (21) 22 (12) 21 (11) 14 (7) 189 (100) 2004 25 (12) 77 (36) 5 (2) 8 (4) 40 (19) 10 (5) 33 (15) 15 7) 213 (100) 2005 34 (13) 76 (28) 8 (3) 4 (1) 87 (32) 24 (9) 21 (8) 16 (6) 270 (100) 2006 27 (12) 71 (31) 6 (3) 10 (4) 48 (21) 12 (5) 34 (15) 19 (8) 227 (100) 2007 33 (11) 82 (28) 8 (3) 13 (4) 74 (25) 11 (4) 40 (14) 32 (11) 293 (100) 2008 25 (10) 79 (31) 11 (4) 10 (4) 61 (24) 14 (6) 30 (12) 21 (8) 251 (100) Total 216 1,912 (11) 537 (28) 71 (4) 63 (3) 513 (27) 126 (7) 251 (13) 135 (7) (100) Notes: 1 Figures in parentheses indicate the % of the papers on an ecosystem type of the total published in that year. 2 Entries in bold indicate when the number of papers on an ecosystem type exceeds 20% of the total number of restoration-related papers for that year.

By far the majority of the papers (65%) reported on studies utilising active restoration methods as opposed to passive ones (Table 2.5) but there were no distinct trend to indicate either a strong bias towards or against active restoration. Again, the total number of observations can exceed 1,575 since there are cases in which both passive and active measures were applied

The number of papers by restoration approach is cross-tabulated with whether the study lists an instrument-based monitoring tool or not (Table2.6). Those using active restoration approaches and using instrumental monitoring tools comprise slightly more than half at 839 or 51,7% of the total observations. The high prevalence of instrumental measures of restoration is due to most papers reporting on specific restoration projects. There were only a few assessments of a range of projects or social assessments. This supports our hypothesis about the disconnect between restoration researchers and practitioners as well as between the theory and practice and the investigators and supposed beneficiaries of the restoration. We return to this again in Section 2.5.

Table 2.5 Number of papers by restoration approach

Active Passive Not specified Total 2000 83 (70) 15 (13) 21 (18) 119 (100) 2001 70 (65) 7 (6) 31 (29) 108 (100) 2002 95 (66) 15 (10) 33 (23) 143 (100) 2003 103 (64) 15 (9) 43 (27) 161 (100) 2004 130 (72) 11 (6) 40 (22) 181 (100) 2005 146 (63) 11 (5) 74 (32) 231 (100) 2006 144 (69) 14 (7) 51 (24) 209 (100) 2007 162 (64) 31 (12) 60 (24) 253 (100) 2008 122 (56) 22 (10) 74 (34) 218 (100) Total 1,055 (65) 141 (9) 427 (26) 1,623 (100) Note: 1 Figures in parentheses indicate the % of papers on a given restoration approach of the total published in the same year.

Table 2.6 Number of papers by restoration approach and the use of instrumentation

Active Passive Not Total specified Use of monitoring instruments 839 (51,7) 116 (7,2) 216 (13,3) 1,171 (72,2) No use of monitoring 216 (13,3) 25 (1,5) 211 (13,0) 452 (27,8) instruments Total 1,055 (65,0) 141 (8,7) 427 (26,3) 1,623 (100) 1 Figures in parentheses indicate the % of papers by category.

Combined, 38% of the studies used re-seeding, planting and/or succession as the primary method of restoration (Figure 2.1a). The alternative restoration approaches include alien vegetation clearing, bush removal, herbivore exclusion, restoration of habitats, wetland restoration, soil excavation and movement, soil fertilization, fire management, change in agricultural practices, and hydrological manipulations.

Figure 2.1 The restoration method utilised, ecosystem function and the constituents of well- being being addressed or affected We also analysed the motivation for engaging a restoration project against the main categories of the ecosystem functions as defined in the Millennium Ecosystem Assessment (MA 2005). Cultural, provisioning and regulating services share the bulk of the research focus (Figure 2.1b). It should be noted that many of the restoration activities were designed to ‘restore’ biodiversity or ‘restore’ plant or animal populations. As there is insufficient evidence to show that such restoration efforts would provide material, supporting or regulating services, by default they were included under cultural services. This is reasonable because society, by and large, is concerned with conserving diversity and considers threatened, or otherwise ‘special’ species, as worth preserving since they are part of our natural heritage or patrimony. This might imply, however, that ecologists may be prioritising the natural good (biodiversity conservation directly) rather than putting it in the context of the social good; on this subject a distinction is only very rarely made explicit. Further support for this argument is provided by the fact that more than 60% of the studies can be classified as contributing towards improved social relations and security, while only a quarter contribute to the provision of materials (Figure 2.1c).

2.4.2 Comparative analysis: Restoration Ecology and the other 12 journals

Restoration Ecology is arguably the world’s leading journal in the field of ecological restoration. We therefore asked the question: Does the profile of restoration related papers published by Restoration Ecology differ from those published by other ecology, agriculture or conservation-

oriented journals? To take this argument further, Restoration Ecology is the journal of the Society for Ecological Restoration (SER) so it is likely that most Restoration Ecology authors either agree, support or strongly sympathise, with the SER’s Primer on Ecological Restoration that defines ecological restoration as “the process of assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed” (SER 2002), i.e. putting a degraded ecosystem on a trajectory of recovery. The SER further stipulates the following specific working definitions and operational concepts – among others – with regard to the attributes of a “restored” ecosystem:

• The restored ecosystem contains a characteristic assemblage of the species that occur in the reference ecosystem and that provide appropriate community structure. • The restored ecosystem consists of indigenous species to the greatest practicable extent. • All functional groups necessary for the continued development and/or stability of the restored ecosystem are represented. • The physical environment of the restored ecosystem is capable of sustaining reproducing populations of the species necessary for its continued stability or development along the desired trajectory. From this partial but representative list, especially the last item, it is evident that the SER’s Primer focuses on ecological concerns, notably on the rebuilding or reconstitution of indigenous plant and animal assemblages, and the repairing of disturbed ecosystem processes, within a landscape perspective. No socio-economic or cultural attributes of “restored” ecosystems are explicitly addressed in detail in the SER Primer, even though they are explicitly mentioned in the Society’s mission statement (see www.ser.org). It is unlikely that authors in other journals – except for Ecological Restoration, the non-peer reviewed journal of SER (not included in this study) – will have this specific focus. Furthermore, it seems logical that Restoration Ecology authors (and readers) are already involved in, and committed to the idea of, restoration and thus do not need to be convinced that restoration is useful or has value, be it monetary of otherwise. These authors are, for the most part, it seems, interested in developing new and improved techniques for doing what they believe needs to be done as well as understanding the underlying ecology. Readers of, and contributors to, other journals may share this focus, but will also be involved in other areas of ecology, and other approaches to restoration, not to mention conservation and/or economic development activities other than restoration, such as ecological engineering, industrial ecology, agro-forestry, sustainable agriculture, etc. The corresponding scientific journals are therefore likely to focus more on people, conservation in general, urban or industrial waste and pollution, agriculture, or other related matters. They are, in sum, likely to be more interested in ecological services to people, and the costs and benefits of restoration than with techniques of restoration as applied to ecosystem functioning and self-organisation per se, or the details of species composition, or a given species’ status, for example, within ecosystems undergoing projects aiming at restoration.

Does the comparative analysis of the “hits” support these assertions? As expected, the percentage of “hits” for Restoration Ecology, on average 87,2% for the entire period, far exceeds that of the other journals which have an average of 5,6% (Table 2.7). Interestingly though, there is a strong increase in the number and percentage of hits over the period 2000 to 2008 for Restoration Ecology, a trend that is not present in the other journals.

Table 2.7 Comparison between Restoration Ecology and 12 other journals

Restoration Ecology Other 12 journals 2003- 2006- 2000- 2003- 2006- 2000-2002 2005 2008 Total 2002 2005 2008 Total Number of papers1 164 204 228 596 4,891 6,426 7,634 18,951 Number of “hits”2 132 176 212 520 221 388 446 1,055 % hits of total3 80.5% 86.3% 93.0% 87.2% 4.5% 6.0% 5.8% 5.6% % PES of “hits”4 2.3% 4.0% 1.4% 2.5% 11.3% 9.5% 11.0% 10.5% % Agric link of “hits”5 54.5% 32.4% 30.2% 37.1% 33.9% 36.6% 38.6% 36.9% % no or little policy outcome of “hits”6 18.2% 64.8% 72.2% 56.0% 93.7% 91.8% 94.8% 93.5% Note: 1 The total number of papers published by either Restoration Ecology or the other 12 journals during the various years. 2 The number of papers that was identified as “hits”, i.e., containing either the word restoration or rehabilitation in the title, abstract and/or keywords. 3 The number of “hits” as a percentage of the total number of papers published. 4 The percentage of the “hits” that made explicit reference to a form or system of payment, or that perceives the possible formation of a payment or system, for ecosystem goods and services. 5 The percentage of the “hits” that indicate an explicit link to agriculture and agricultural productive systems. 6 The percentage of the “hits” that has no impact, or only a minor impact, on policy.

One significant difference between the two sets is that only about 2,5% of the “hits” in Restoration Ecology, address or refer to payments for ecosystem goods and services. The period 2006-2008 was the lowest with only 1,4% explicitly dealing with PES. In contrast, the other 12 journals averaged 10,5%, a four-fold difference over the entire period. Clearly, authors in Restoration Ecology are not focusing on restoration projects or programmes that are explicitly linked to the introduction of payment systems. This is further emphasised by the fact that 54% of the studies were authored by academics or researchers attached to scientific institutions, and 17% by researchers in government agencies, (Figure 2.2) whereas only about 11% were submitted by private companies, land owners and community collectives combined.

Figure 2.2 Institutional demarcation of those conducting restoration projects as published in Restoration Ecology Another marked difference between the journals is whether the published papers on restoration have any link to agriculture. While the percentage of papers (“hits”) with an agriculture link increased over time from 34% to almost 39%, for the 12 journals other than Restoration Ecology, declined sharply over time from a high of 55%, in 2000-2002, to about 30%, in 2006-2008. As expected and as discussed above, there are therefore clear and marked differences between the two sets of papers. It should be noted that most of the papers dealing with agriculture or indicating a link with the agricultural sector, focus on the effects of restoring degraded, abandoned agricultural lands, also known as old fields, in formerly woody vegetation types. It is therefore an attempt to reintroduce an original or different, nature-based, land-use option after failed or abandoned agricultural production. The cost of not managing farmland or of farming in ways in environmentally- benign ways is therefore a delayed cost and a cost deferred to other parties.

From a socio-political vantage point, the increase in papers in Restoration Ecology with little impact on policy outcome, and the high level of similar papers in the other 12 journals, is a concern. It does seem, however, as if Restoration Ecology still has a marginally stronger policy focus than the other 12 journals. Matters are aggravated by the fact that many of the papers only refer to plausible policy impacts, but in and by themselves do not state actual policy changes or impacts. Most simply make recommendations for improved or best practice and do not actually mention or discuss any specific policies.

Table 2.8 Comparison between Restoration Ecology and 12 other journals concerning the number of papers by scale1

Restoration Ecology 12 Journals Landscape Ecosystem Community Population Landscape Ecosystem Community Population 2000 9 (26) 26 (76) 3 (9) 1 (3) 43 (56) 28 (36) 17 (22) 7 (9) 2001 16 (44) 14 (39) 3 (8) 0 (0) 42 (58) 34 (47) 32 (44) 12 (16) 2002 38 (61) 22 (35) 1 (2) 1 (2) 32 (45) 31 (44) 13 (18) 13 (18) 2003 13 (26 28 (56) 8 (16) 0 (0) 55 (53) 54 (52) 26 (25) 8 (8) 2004 11 (19) 31 (54) 15 (26) 1 (2) 60 (48) 54 (43) 31 (25) 18 (14) 2005 22 (32) 18 (26) 32 (46) 4 (6) 79 (50) 88 (55) 40 (25) 11 (7) 2006 19 (28) 36 (53) 9 (13) 3 (4) 70 (51) 70 (51) 43 (31) 9 (7) 2007 44 (46) 38 (40) 36 (38) 16 (17) 68 (48) 65 (45) 39 (27) 21 (15) 2008 26 (54) 5 (10) 8 (17) 9 (19) 74 (45) 54 (33) 38 (23) 31 (19) Total 198 (35) 218 (39) 115 (20) 35 (6) 523 (37) 478 (34) 279 (20) 130 (9) Note: 1 Figures in parentheses indicate the % of the number of papers for a given scale relative to the total for that year.

Although there is no meaningful trend over time for either Restoration Ecology or the other journals, it is evident that restoration-related papers in both Restoration Ecology and the 12 other journals scrutinized are dominated by those focusing on landscape and ecosystem restoration projects (Table 2.8). For Restoration Ecology, 74% of all papers included in this study fall into this category, while the corresponding value for the other 12 journals combined is 71%.

One explanation for this finding is that the majority of the studies deal with restoration in forest, savanna and woodlands, or in rivers and wetlands (Table 2.4) where a focus on landscape or ecosystem scale is to be expected. There is therefore a clear focus on interacting biotic communities and the processes that operate across larger areas. The same bias is, not surprisingly, found in the relationship between scale and PES (Table 2.9). More than 80% of all PES cases are reported for restoration on a landscape and ecosystem scale. This is to be expected since, due to PES’s high transaction cost, it is unlikely that single species or single communities would provide sufficient value to justify development of a payment system. Also, from a PES perspective, there is less of an emphasis on species and more on restoring functioning ecosystems and the diversity of communities that support functioning landscapes (e.g. restoring catchment water storage through soil retention measures) which provide benefits to people across the full spectrum of society.

Table 2.9 Comparison between Restoration Ecology and 12 other journals concerning PES and scale

% of PES observations Number of observations Landscape Ecosystem Community Population Restoration Ecology 53% 27% 13% 0% 14 12 Journals 43% 40% 12% 6% 121

2.4.3 Relative state of national economic development and restoration

Natural capital provides a flow of ecosystem goods and services essential to life (Costanza et al. 1997; Daily 1997, De Groot et al. 2002, MA 2005). This implies that intact or restored natural capital is very important from an economic development perspective (Aronson et al. 2007), especially when it is recognized that some of the main causes for environmental degradation are rangeland grazing and the harvesting of fuelwood at rates faster than replacement by primary production is possible (Duraiappahn 1998, Mahiri and Howorth 2001, Ayyad 2003, Wezel and Bender 2004, Geerken and Ilaiwi 2004, Wessels et al. 2004). Does the importance of intact of restored natural capital correspond to the level of economic development of a country, or is restoration mainly the focus of the relatively rich?

To address this question for the data assembled in this study, we used the income classification system of the World Bank as listed in its Word Development report of 2007 (World Bank 2006) to categorise the countries in which the restoration took place. By far the majority (72%) of the 1,575 “hits” were studies reporting on restoration in high income countries with the USA and Europe dominating the list (Figure 2.3). Only 3% of all the studies (42 in total) were conducted in low income countries with India and central and eastern Africa (Ethiopia, Kenya, Tanzania and Uganda) comprising about 80% of all the country studies in this income group. While the need for restoration from a livelihoods perspective is high in these countries, they do not have the resources or the people to actually conduct the restoration and report on it in academic journals.

Figure 2.3 Distribution of studies by the economic status of the host country The demand for restoration from a livelihoods perspective is highlighted by the fact that 65% of the studies in low income countries and 67% of the studies in low middle income countries focuses

on provisioning and regulating services (Table 2.10) (the number of observations can exceed 1,575 due to the fact that restoration can address various functions simultaneously). The corresponding figure for high income countries is 55%. High income countries also have, proportionately, a considerably higher number of studies focusing on cultural services than the low income countries. This may reflect two restoration drivers that are more important in high income countries: (a) restoration of threatened or otherwise special species; and (b) restoration aimed at restoring the recreational (use or non-use) or aesthetic and amenity value of an environment. Both of these were classified as cultural services. This is a clear indication that the focus and the preference and the need or objective of the restoration activity differs among countries with a different income profile.

Table 2.10 Distribution of studies by ecosystem function in terms of the economic development status of the host country

Low income Low middle Upper High income Not Total income middle specified or income others Supporting 13 (3%) 23 (5%) 23 (5%) 298 (70%) 67 (16%) 424 (100%) (16%) (13%) (12%) (16%) (21%) (16%) Regulating 19 (2%) 65 (8%) 58 (7%) 557 (71%) 86 (11%) 785 (100%) (23%) (36%) (31%) (30%) (26%) (30%) Provisioning 34 (5%) 57 (8%) 58 (8%) 466 (67%) 83 (12%) 698 (100%) (42%) (31%) (31%) (25%) (25%) (27%) Cultural 15 (2%) 37 (5%) 51 (7%) 523 (73%) 90 (13%) 716 (100%) (19%) (20%) (27%) (28%) (28%) (26%) Total 81 (3%) 182 (7%) 190 (7%) 1,844 (70%) 326 (12%) 2,623 (100% (100%) (100%) (100%) (100%) (100%) (100%) Note:

1 Figures in parentheses in the same line as the number of observations indicate the % of the observations by income category relative to the total number of observations. The figures in parentheses below the number of observations indicate the percentage distribution within each income category.

When considering the studies’ contribution to human well-being, the same trend is observed (Table 2.11). High income countries focus much less attention on provisioning services (24%) than low income countries (38%), but much more on social relationships (30%) compared to low income countries (21%). This illustrates the focus and the need of restoration in low income countries to be on aspects related to livelihoods and to a lesser extent other welfare issues. No noteworthy trend or observation, however, emerges in terms of the distribution of policy outcome when cross-tabulated with the income level of the host country (Table 2.12).

Table 2.11 Distribution of studies by contribution to human well-being in terms of the economic development status of the host country

Low Low Upper High Not Total income middle middle income specified income income or others Material 30 (5%) 49 (8%) 49 (8%) 429 (66%) 93 (14%) 650 (100%) (38%) (26%) (27%) (24%) (29%) (25%) Health 8 (2%) 27 (7%) 27 (7%) 263 (69%) 58 (15%) 383 (100%) (10%) (15%) (15%) (15%) (18%) (15%) Security 24 (3%) 69 (9%) 56 (7%) 564 (71%) 80 (10%) 793 (100%) (31%) (37%) (31%) (32%) (25%) (31%) Social 16 (2%) 41 (6%) 50 (7%) 532 (73%) 89 (12%) 728 (100%) relations (21%) (22%) (27%) (30%) (28%) (29%) Total 78 (3%) 186 (7%) 182 (7%) 1,788 320 (13%) 2,554 (100%) (100%) (100%) (100%) (70%) (100%) (100%) (100%) Note:

Table 2.12 Distribution of studies by policy outcome in terms of the economic development status of the host country

Low Low middle Upper High Not specified Total income income middle income or others income None 31 (3%) 76 (7%) 82 (7%) 804 112 (10%) 1,105 (100%) (72%) (68%) (74%) (73%) (67%) (70%) (70%) Local 7 (3%) 19 (7%) 20 (7%) 212 16 (6%) 274 (100%) (16%) (17%) (18%) (77%) (10%) (17%) (19%) National 4 (3%) 12 (10%) 5 (4%) 82 (68%) 18 (15%) 121 (100%) (9%) (11%) (5%) (7%) (11%) (8%) Global 1 (1%) 5 (6%) 4 (5%) 46 (60%) 21 (27%) 77 (100%) (2%) (4%) (4%) (4%) (13%) (5%) Total 43 (3%) 112 (7%) 111 (7%) 1,144 167 (11%) 1,577 (100%) (100%) (100%) (100%) (73%) (100%) (100%) (100%) Note:

What is truly encouraging, remarkable, and unexpected, is that there is a bias towards an agricultural link among poor countries (Table 2.13). While only 35% of the papers from high-income countries indicate an agriculture link, 57% of the low income countries do indicate such a link with 43% for low middle income and 52% of upper middle income countries. This clearly indicates that among the poorer countries the link to productive agricultural systems is important and there is recognition that restoration can add value to agricultural systems. As it is well-known that agriculture is the dominant economic activity among poorer nations, it is good to see that the emphasis, in practise, of restoration is in support of this sector.

Table 2.13 Distribution of studies with an agriculture link in terms of the economic development status of the host country

Low Low Upper High income Not specified Total income middle middle or others income income Agric 24 (4%) 48 (8%) 58 (5%) 406 (70%) 46 (12%) 582 (100% Yes (57%) (43%) (52%) (36%) (27%) (37%) Agric 18 (2%) 63 (6%) 53 (5%) 737 (74%) 122 (12%) 993 (100%) No (43%) (57%) (48%) (64%) (73%) (63%) Total 42 (3%) 111 (7%) 111 (7%) 1,143 (73%) 168 (11%) 1,575 (100%) (100%) (100%) (100%) (100% (100%) (100%) Note:

An interesting observation is that the upper middle income countries are driving the process to link restoration with markets and payments for ecosystem goods and services (Table 2.14). A total of 12% of all studies from upper middle income countries do specify a PES link, which is almost double that found among studies from high income countries (7%). Upper-middle income countries that feature strongly are Mexico, Costa Rica, Poland, South Africa and the Czech Republic.

Table 2.14 Distribution of studies with a PES link in terms of the economic development status of the host country

Low Low middle Upper High income Not Total income income middle specified income or others PES Yes 2 (2%) 7 (6%) 13 (10%) 84 (68%) 18 (15%) 124 (100%) (5%) (6%) (12%) (7%) (11%) (8%) PES No 40 (3%) 104 (7%) 98 (7%) 1,059 (73%) 150 (10%) 1,451 (95%) (94%) (88%) (93%) (89%) (100%) (92%) Total 42 (3%) 111 (7%) 111 (7%) 1,143 (73%) 168 (11%) 1,575 (100%) (100%) (100%) (100%) (100%) (100%) (100%) Note:

From the above analysis it is clear that economic development status matters. It not only matters whether people in a given country engage in and report on restoration in the scientific literature, it also matters in terms of project objective and purpose as well as the project’s link to agriculture and PES. The emphasis on basic, livelihood-linked services in the papers studied decreases as one goes from poor to rich host countries, and conversely, the importance of a social and cultural focus increases. Poorer nations also have a stronger focus on agriculture in these papers dealing with ecological restoration. The surprising result, however, is the link to PES, which is a process driven by mainly upper-middle income countries.

2.4.4 Restoration, payments for ecosystems, agriculture and impact on policy

Out of the 1,575 papers examined, only 582 (37%) indicated a link with agriculture (Table 2.15). There is also a significant difference between the “yes” and “no” results since the Pearson’s Chi2- statistic is 117,3 (df = 10 & P < 0.00001). This relatively weak link is in marked contrast to the substantial contribution of restoration towards cultural and material well-being (from a functional perspective) and toward material goods and social relationships from a well-being point of view (Figure 2.1). The link to agriculture is, however, strongest in papers dealing with grasslands, shrublands and urban and transformed ecosystems. It is projects undertaken in sites corresponding to these three ecosystem types that the number of “yes” observations – indicating a link between restoration and agriculture – exceeds the number of “no” observations (Table 2.10). However, papers on these ecosystems comprise only a small proportion of the whole, and the bulk of the papers (55%) are either on forest, woodland and savannas or rivers and wetlands (Table 2.4). These two broad groups of ecosystems also comprised 48% of the total number of “yes” observations. The skewed distribution is to be expected because grasslands are a highly productive ecosystem type

and are heavily used for grazing systems as well as being cultivated for the production of crops. The intensive use made of grasslands also renders them vulnerable to degradation and, thus, also in need of restoration. By contrast, arid and semi-arid, and coastal, ecosystems generally are relatively unsuitable for agriculture.

Table 2.15 Agricultural link with restoration projects in varying types of ecosystems

Link to Grass- Forests, Shrub- Arid and Rivers Marine and Urban and Unclassi- Total Agriculture lands woodlands lands semi-arid and coastal transformed fied and savanna ecosystems wetlands areas Yes 137 (19) 192 (26) 41 (6) 23 (3) 162 (22) 30 (4) 117 (16) 34 (5) 736 (100) No 79 (7) 345 (29) 30 (3) 40 (3) 351 (30) 96 (8) 134 (11) 101 (9) 1,176 (100) Total 216 (11) 537 (28) 71 (4) 63 (3) 513 (27) 126 (7) 251 (13) 135 (7) 1,912 (100) Notes:

1 Figures in parentheses in the same row as the number of observations indicate the % distribution of the number of papers with or without an agricultural link by ecosystem.

2 Figures in bold indicate those ecosystem types in which the % “yes” exceeds the % “no”.

The extremely low number of restoration-related studies that deal with the development of markets for ecosystem services is a cause for concern (Table 2.16). As indicated in Section 1, markets are social constructs that reflect the will and the intent of the people. If there is no clear and discernable link between restoration and PES or the establishment of markets, it derives from the fact that restoration is not yet viewed by society at large as an activity that reflects its desires, needs or wishes. Only 124 of 1,575 (8%), i.e., 0.6% of the restoration studies examined addressed the topics of markets or payment systems for ecosystem services. The proportion of studies that did examine PES systems has not changed over the last 8 years, which is surprising given the increase in interest in ecosystem services in other fields such as conservation planning. This is a clear indication that there is a huge gap in the ecological restoration literature on the subject of development of PES markets.

Table 2.16 The development of markets for ecosystem services for restoration-related activities

Yes No Total If yes, has a payment system been formed or only been perceived Formed Perceived Total 2000 12 (11) 99 (89) 111 (100) 12 (86) 2 (14) 14 (100) 2001 7 (6) 102 (94) 109 (100) 4 (100) 0 (0) 4 (100) 2002 9 (7) 124 (93) 133 (100) 5 (56) 4 (44) 9 (100) 2003 13 (8) 141 (92) 154 (100) 11 (85) 2 (15) 13 (100) 2004 11 (6) 171 (94) 182 (100) 7 (70) 3 (30) 10 (100) 2005 20 (9) 208 (91) 228 (100) 16 (80) 4 (20) 20 (100) 2006 19 (9) 186 (91) 205 (100) 16 (80) 4 (20) 20 (100) 2007 13 (5) 226 (95) 239 (100) 11 (79) 3 (21) 14 (100) 2008 20 (9) 194 (91) 214 (100) 18 (86) 3 (14) 21 (100) Total 124 (8) 1451 (92) 1,575 (100) 100 (80) 25 (20) 125 (100)

The strong link between restoration and agriculture is not surprising given the fact that 30 of the studies were concerned with markets in the forest, woodlands and savannas, and 28 with aquatic ecosystems (rivers and wetlands) (Table 2.17). Together they represent 42% of the total studies where PES were evoked or specified. The strong emphasis on woody plant-dominated ecosystems is to be expected given their well-recognised importance for carbon sequestration and known impacts on water quantity, quality and flow regulation. The difference between the “yes” and the “no” is statistically significant since the Pearson’s Chi2-statistic is 27,1 (df = 10 & P=0,0025). The undeveloped nature of the market for ecosystem services will be discussed in detail in Section 5.

Table 2.17 Markets for ecosystem services by ecosystem type

Link to Grass- Forests, Shrub- Arid and Rivers Marine Urban Unclassi- Total payment lands woodlands lands semi-arid and and and fied systems and ecosystems wetlands coastal trans- savanna areas formed Yes 19 (14) 30 (22) 7 (5) 2 (1) 28 (20) 7 (5) 32 (23) 12 (9) 137 (100) No 1,775 197 (11) 507 (29) 64 (4) 61 (3) 485 (27) 119 (7) 219 (12) 123 (7) (100) Total 1,912 216 (11) 537 (28) 71 (4) 63 (3) 513 (27) 126 (7) 251 (13) 135 (7) (100) Notes:

1 Figures in parentheses in the same row as the observation indicate the % distribution of the number of papers with or without a link to markets for ecosystem services by ecosystem type.

3 Figures in bold indicate those ecosystem types in which the % “yes” exceeds the % “no”.

When the link between agriculture and PES is cross-tabulated (Table 2.18) it is evident that only 5,5% of all the papers have both a PES and an agriculture link. This stands in stark contrast with the 61% of papers that have neither a PES nor an agriculture link and are, therefore, purely biophysical and “ecology” oriented. These findings clearly indicate that there is a strong bias, at least in the

peer-reviewed academic literature, to keep restoration research focused on ecological science and not on the economic and social aspects and implications. This is a missed opportunity because the link between restoration and economic development and social welfare is a strong and direct one (Aronson et al. 2007). A focus only on the biophysical reasons for restoration is unlikely to lead to the development of more opportunities for restoration. There is also significantly more “no Agric” and “no PES” papers than “PES” and “no PES” but with agriculture, as indicated by a statistically significant Yates continuity corrected Chi2-statistic of 62.2 (df = 1 & P<0.0001).

Table 2.18 Comparison between PES and agriculture

PES No PES Total Agric 87 (5,5) 495 (31,4) 582 (37,0) No agric 37 (2,3) 956 (60.7) 993 (63,0) Total 124 (7,9) 1451 (92,1) 1575 (100) Note:

1 Figures in parentheses indicate the % distribution of the number of papers per category in terms of the total number of “hits”.

Among the 582 restoration-related papers that do indicate a link with agriculture, 17% are aimed at improving agriculture, 28% focus on changing from one form of agricultural land-use to another (less damaging) form of land use, and 34% focus on removing the effects of, or the scars left by, agriculture. A further 21% of the papers provide lessons for the application of restoration in an agricultural context (Table 2.19). The entries in Table 2.19 are listed by journal and sorted in a descending fashion in terms of the expected improvement in agriculture. It is interesting to note that the Journal of Arid Environments, which deals with regions that are generally unproductive and poorly developed and often degraded, tops the list. Most restoration papers here relate to agricultural improvement, such as grazing land restoration to provide better grazing. This is a clear indication of the importance of improving agriculture under arid conditions and the contribution restoration can make towards this objective. This is in stark contrast to, for example, the more Euro- centric journal Agriculture Ecosystems and Environment in which most agriculturally linked restoration papers focus on removal of the impacts of agriculture, i.e. the conversion of arable and grazing land to natural vegetation. It is also worth noting that some journals focused strongly on change in agricultural practices – particularly modifying the way in which agricultural landscapes were managed so as to reach a compromise between economic values (money earned by crops and livestock) and other values (aesthetics, biodiversity conservation, flood regulation, environmental health, security). Journals that appeared to have this focus for restoration are Environmental Development Economics, Conservation Biology, Ecological Economics, Journal of Applied Ecology and Frontiers in Ecology and Environment.

Table 2.19 Treatment of agriculture related to restoration in various journals

Journal A link with Percentage distribution of those where a link to agriculture agriculture is described1 described Yes No Improve Change Remove Lessons agric agric agric Journal of Arid 42 18 74% 12% 14% 0% Environments Environmental 2 0 50% 50% 0% 0% Development Economics Journal of Environmental 8 35 38% 25% 13% 25% Management Restoration Ecology 193 327 21% 21% 44% 15% Agriculture, Ecosystems 29 5 17% 28% 62% 0% and Environment Ecological Engineering 56 163 14% 38% 30% 21% Conservation Biology 13 35 8% 62% 23% 8% Journal of Applied 43 89 7% 49% 12% 33% Ecology Forest Ecology and 72 188 6% 25% 33% 36% Management Biological Conservation 97 90 4% 27% 32% 38% Ecological Economics 17 10 0% 59% 41% 0% Water SA 5 0 0% 20% 20% 60% Frontiers in Ecology and 5 33 0% 80% 20% 0% Environment Total 582 993 17% 28% 34% 21% Note:

1 It is possible for the % to exceed 100 since multiple entries are possible.

Apart from readership, journal policy also biases the papers. For example, Restoration Ecology focuses on the conversion of agricultural land to set-aside natural land (removal of agriculture). The other 12 journals, however, focussed significantly more on the contribution that restoration could make to improving, or changing agriculture, or that agriculture could make to informing restoration (Table 2.20, Chi Square (with Yates continuity correction) = 11.7, df = 1 & p<0.001). This significant distinction is a reflection on the differing editorial policies as reflected in Table 2.20 where Restoration Ecology’s explicit objective is ecosystem recovery after degradation.

Table 2.20 Different treatments of agriculture between Restoration Ecology and the other 12 journals

Improve or change agric or Remove lessons from agriculture Total Other 12 journals 280 114 394 Restoration Ecology 109 84 193 Total 389 198 587

Table 2.21 Monitoring tools used during ecosystem restoration

Year Monitoring Monitoring Total Monitoring tools applied done not done Instrumental Interviews Others Total 2000 74 (67) 37 (33) 111 (100) 61 (71) 4 (5) 21 (24) 86 (100) 2001 72 (66) 37 (34) 109 (100) 67 (91) 3 (4) 4 (5) 74(100) 2002 108 (82) 24 (18) 132 (100) 99 (86) 3 (3) 13 (11) 115 (100) 2003 122 (79) 32 (21) 154 (100) 115 (94) 2 (2) 5 (4) 122 (100) 2004 113 (62) 69 (38) 182 (100) 104 (90) 1 (1) 11 (9) 116 (100) 2005 138 (61) 90 (39) 228 (100) 111 (79) 1 (1) 28 (20) 140 (100) 2006 153 (75) 52 (25) 205 (100) 149 (96) 4 (3) 2 (1) 155 (100) 2007 180 (75) 59 (25) 239 (100) 159 (84) 14 (7) 16 (8) 189 (100) 2008 170 (79) 44 (21) 214 (100) 160 (92) 4 (2) 10 (6) 174 (100) Total 1,130 (72) 444 (28) 1,574 (100) 1,025 (88) 36 (3) 110 (9) 1,171 (100)

By far the majority of the studies (72%) discussed specific monitoring tools, and 88% utilised some form of instrumental measurement (Table 2.21). This may, to a certain extent, indicate the relative ease of doing basic physical measurements or (visual) assessments above say complex surveys and stakeholder participatory-based research. With the measurement there is also a degree of scientific rigour concerning restoration efforts, without which it would have been difficult to compare and replicate analyses. It also, unfortunately, indicates a strong biophysical bias and a very low degree of stakeholder involvement – people are not asked and/or engaged in the process. Only 3% made use of interviews and questionnaires or surveys despite the fact that restoration is perceived as making a considerable contribution to human well-being, social relations and cultural values. Scientists are, therefore, clearly operating in an academic vacuum while conducting their research and simply indicating that it will have “some effect” that will be perceived by society as a benefit. This assumption is, however, seldom tested. There may, therefore, be a disconnect between the technical successes of a restoration project and the actual benefits that it is perceived to deliver. A plausible reason for this may be that this social interaction could be perceived as “outside the realm of science” – and that such issues are best relegated to the grey literature. With the academic reward system based on citation indices, and peer review, and not on the impact on implementation and practice of restoration, there may be little incentive for academics to do otherwise.

Figure 2.4 The indicated policy outcome (a) and intensity (b) of restoration-related studies The divide between researchers and the rest of society also is substantiated by the fact that between 70% and 80% (and rising) of the papers indicated no policy outcome (Figure 2.4a). Policy outcome is measured in terms of the effect of the study on policy at various political levels. The same increase is evident in the low policy intensity ratings (Figure 2.4b) which are between 60% and 70%), where intensity is a measure of the number of people affected by the policy or the importance of the policy for the way in which the world is run. Also, in most cases where reference is made to policy outcomes, it is the researchers’ own expressions and recommendations towards policy, but there is no indication of the real impact a study has, or might have, on policy.

What we do not know from this literature review is the extent to which scientific knowledge is transferred to users in a different way, for, by example, books, the grey literature, participation in workshops and seminars, and/or other teaching material. While it is true to say that practitioners and managers and policy-makers generally find academic papers difficult to read, it is also true to say that knowledge has a short life span if it is not published and applied. Clearly there is a dual responsibility to publish research outputs and to communicate them in other ways as well.

2.5 Discussion

As stated above, we studied 19,547 papers published between January 2000 and October 2008 in 13 peer-reviewed academic journals. Of this total, only 1,575 (8%) contains the words restoration or rehabilitation in their titles, abstracts or keywords. Only 124 (8%) of the 1,575 “hits”, or 0.6% of the total number of papers examined, make reference to the establishment of a payment system or a market for ecosystem services. These are seriously worrying statistics that also signal an opportunity to develop new ground. Matters are aggravated by the fact that 61% of all the hits – i.e. restoration or rehabilitation related papers, made no reference at all to either a payment system or to links with agriculture. Restoration research, as reflected in the leading journals, is therefore conducted in virtual isolation from its value and contribution to people. This reasoning is supported by the fact that while 72% of the papers stated an explicit monitoring and evaluation tool, 88% of them use an instrumental measure and only 3% interviewed people. This finding emphasises the disconnect between people and restoration activities and is also supported by the fact that 80% of the papers do not have any discussion or analysis of direct policy impact or implication. There is therefore a highly noteworthy disconnect between society, and the economy in general, and restoration ecology and restoration activities. For restoration to be taken seriously in future, this disconnect will have to be addressed.

Returning to the purpose of this study, which was to develop an understanding of the use of monitoring and evaluation in restoration programmes and its implication for the development of markets for ecosystem goods and services, we have to conclude the following:

• That the concepts of ecosystem services (i.e. explicitly linking services to beneficiaries and demonstrating values) has not yet entered the mainstream academic press insofar as it reports on restoration ecology, and therefore could be regarded as still in its infancy in this field. There are at least three reasons for this: o The restoration community has not yet begun to adopt the new ‘paradigm’ of RNC, and employ its language and precepts; therefore even if the services and benefits of restoration are there implicitly, as in the SER’s Mission statement, they are not studied or expressed explicitly or in sufficient detail. o The researcher community involved in ecosystem services are isolated from those in the restoration field. This is typical of the problem of academic sequestration and resistance to truly inter-disciplinary, or trans-disciplinary studies. o Those involved in developing economic development pathways have not yet considered natural capital and the restoration thereof as a catalyst for such development.

• Most papers, and hence the monitoring and evaluation component thereof, are from short- term studies. They are science and research-based rather than project or programme orientated. The monitoring and evaluation reflects the general lack of support for longer- term (>3 yrs) studies. This is seriously disconcerting given the fact that the full benefits of restoration may only become evident after 5 or more years, particularly in arid and semiarid environments. • There is an extremely weak link between policy and research and the policy indications provided are often simply the researchers’ recommendations rather than actual policy impact. • Some of the major gaps in the papers studied relate to the following: o The lack of studies that target the relevant policies/practices and practitioners; o Failure to engage actively with those who are the beneficiaries of the findings. • The lack of development of markets and payments systems in restoration related papers is the most disconcerting aspect. In a high-ranking, highly regarded journal such as Conservation Biology, there has not been a single paper since 2000 which is both a “hit” and deals explicitly with ecosystem services. Among the papers that were hits, only four make direct references to the concept of ecosystem services. In the Journal of Arid Environments there have been only two. • While very few references to funding sources are made, most funding seems to have come from research grants or contracts and thus the sponsors, rather than the beneficiaries, are paying for the research work. • The gaps between researchers and policy, researchers and society-at-large, and restoration and payment systems, should all be addressed as a matter of urgency, given the extremely large potential impact restoration has or can have on society, and the environment. This is further supported by the fact that half of the papers studied do have a strong link to agriculture production systems. The incentive to make this effort at rapprochement is increased by the potential that restoration has to reduce the current and future impacts of climate change and the ongoing global food security crisis.

This study was designed to test the hypothesis that restoration improves water flow and water quality, land productivity, in some instances sequesters more carbon, and, in general, improves both the socio-economic value of the land in and the surroundings of the restoration site as well as the agricultural potential of the land. While this is our hypothesis, and one that should intuitively be considered to hold, there is very little if any evidence from the literature review carried out to date to support it, especially if one adds, to restoration, the further condition of the development of markets for ecosystem services. This unexpected finding reinforces the value of the research project of which this literature study is one small part, and of its potential contribution not only to South African science and policy, but also globally.

Although they are disconcerting, the findings of this study thus far are, unfortunately, in line with results found elsewhere. The growing call for evidence-based conservation research activities during recent years is for a rigorous, well documented, scientific process, both in the way conservation and ecosystem management decisions are taken and in the measurement of their outcomes (Sutherland et al. 2004). There is therefore a definite shift in focus from 'inputs' and

'outputs' to 'outcomes' (Ferraro et al. 2006). Markets react to outcomes rather than to either inputs our outputs per se, irrespective how valuable they might be. It therefore seems as if natural scientists have yet to cross this bridge that would lead to a new social construct – a new market – for restoration-related ecosystem services. They may require the help of social scientists to do so. In the light of this strong call for outcomes it is even more important to acknowledge the human factor in the design of conservation and restoration projects (Saunders et al. 2006, Cowling and Wilhelm- Rechmann 2007). Unfortunately, this “human choice” factor, which is critical to the successful implementation of conservation goals and restoration outcomes, is often neglected (Knight et al. 2006).

2.6 Conclusion

This Chapter presents a comprehensive literature review on the benefits of ecological restoration and rehabilitation for society. The results are based on 19,547 peer reviewed journal papers published between 2000 and 2008 of which 1,575 ‘hits’ contained the words ‘restoration’ or ‘rehabilitation’ in their titles, abstracts and/or keywords. These ‘hits’ were analysed according to a set of criteria measuring to what extent restoration explicitly links to broader societal well-being.

Our review clearly illustrates that the tangible contributions of restoration to society are not yet acknowledged or highlighted in the academic literature on restoration. The monitoring and evaluation of restoration projects are largely instrumental, reflecting a basic scientific bias on specific restoration projects. A very low degree of stakeholder involvement is reported, which highlights the worrisome gap between the technical success of a restoration project and the actual benefits it is perceived to deliver. This bias towards ecology or biodiversity, as opposed to society, as the beneficiary of restoration, is particularly evident in the journal Restoration Ecology, the flagship journal of ecological restoration.

The gap between research on restoration and the rest of society is further substantiated by the observation that most of the papers indicate no policy outcomes. Very little research on restoration acknowledges the use of markets to allocate the benefits of ecosystem flows to different users. Only a very small percentage of papers link restoration to the development of payments for ecosystem services (PES) and most are concentrated in higher income countries. This is largely the result of the simple fact that almost three-quarters of the 1,575 ‘hits’ reported on restoration are papers produced in high income host countries.

When restoration is studied in low income countries, it is more directly linked to aspects of livelihoods, such as the provisioning of materials or regulating services. There is also a clear bias towards an agricultural link with restoration in low income countries, reflecting the dominant contribution of agricultural systems to human well-being in such countries. Overall, the link to agriculture is relatively weak and biased due to a relative underreporting of restoration in grasslands, shrublands and urban and transformed systems, which are more clearly associated with highly productive agricultural activities. More than half of the reported restoration is focussed on forest, woodlands and savannas as well as rivers and wetland ecosystems.

It is recommended that the concept of ecosystem services and the value to society become mainstreamed to a higher degree in the field of restoration ecology, and that the longer term effects of restoration on society be more closely studied. Furthermore, policy impacts of restoration studies need to be made more explicit, and most importantly, the benefits of markets and payment systems for restoration work need to be much more clearly researched and communicated.

CHAPTER 3 RESEARCH METHODOLOGICAL ISSUES WITH RESPECT TO SYSTEMS DYNAMIC MODELLING2

Douglas Crookes and Martin de Wit

3.1 Introduction

In this chapter we describe what a systems dynamics (SD) modelling approach is and how it can be applied to complex and dynamic problems. The systems dynamic approach brings the results together from the separate application of ecological, hydrological and economic sciences to the problem of restoration in an attempt to derive insights on a systems level. The value-added by such an approach is that direct comparison between the results of restoration at various sites is made possible and the temporal dynamics and risk profile of the results is made explicit. By presenting results from a systems dynamics integrative model to decision-makers, one starts to address the dual traps of scientific results being (a) only relevant to certain sites and (b) that such results are limited only to a static interpretation of the situation. By using a systems dynamics modelling approach one gains insights into higher-level systems-wide perspectives, longer-term dynamics and an explicit profile of risk over time.

The structure of this chapter is as follows. Section 3.2 gives an overview of the generic steps followed in SD modelling. Section 3.3 presents an example of an environmental-economic application of SD modelling, section 3.4 presents the steps in the modelling process and section 3.5 an overview of the features of a typical SD model. Section 3.6 concludes the chapter.

3.2 Generic steps in SD modelling

In order to compare modelling frameworks, we need a generic set of steps that can be used to compare these frameworks. Having assessed a number of different problem solving frameworks, we elected to use the generic framework of Robertshaw et al. (1978), since this is a systems methodology that has already been applied to a number of economic tools (e.g. cost benefit analysis). The author identifies a number of questions underpinning an iterative process of problem definition and generating and evaluating alternatives. These questions are as follows (Table 3.1):

2 This chapter is largely based on Crookes (chapter 3, in progress). Modelling the ecological-economic impacts of RNC, with a special focus on water and agriculture, at eight sites in South Africa. Ph.D thesis. Department of Economics. Stellenbosch University.

Table 3.1 Stages in system problem solving

Stage Key questions Defining the problem: What is the problem? What must be accomplished? Who is the decision maker? What is the value system? How will he/she pick among the alternatives? What are the constraints? Generating alternatives: What are the alternatives? How will these alternatives operate under the conditions (constraints) of the problem? How much do they cost? What will they produce?

Evaluating alternatives: Which alternatives do I pick? What are the factors affecting the worth of each alternative? Source: Robertshaw et al., 1978

We now apply these generic steps to our research problem.

3.2.1 Defining the problem

Two categories of decision-maker usually occur. The first is the development facilitator, usually an official in the economic affairs directorate or development planning department. The value system of this decision-maker is usually one of growth promotion and economic efficiency. However, a second category of decision-maker also exists. These decision-makers are encountered in environmental, water or agricultural policy contexts, with a value system focussed on environmental protection. This study is focussed on targeting both these categories of decision-makers, in other words promoting economic development while also addressing environmental concerns. These decision-makers require a toolkit that facilitates the identification of key problems within the system under investigation in an integrated framework that addresses economic, social and environmentally aspects.

3.2.2 Generating alternatives

Grant et al. (1997) compare hard system methods such as system dynamics modelling with soft systems modelling, physics and statistics based on two dimensions. The first is the degree of interrelatedness of the components and the second is approaches based on the availability of data and level of understanding of the system (Table 3.2).

Table 3.2 Comparisons between methods of problem solving

Level of Understanding Low High Availability of data High Statistics Physics Low Soft systems analysis and System dynamics modelling simulation Source: Based on Grant et al., 1997.

Traditional economic modelling approaches, it is suggested, fall primarily in the same category as statistics and physics, characterised by a high degree of data availability. However, more often than not when modelling environmental-social-economic systems data are parsimonious (in particular time series data) and approaches need to be developed that take this into consideration.

Boulanger and Brechet (2005) highlight a number of sustainability problems, many of which are of relevance to the current study. These sustainability problems imply a methodological answer (Table 3.3). The authors identify six modelling approaches relevant to assessing these categories of problems. These include multi-agent modelling, system dynamics, Bayesian networks, optimisation, general equilibrium modelling (GEM) and econometrics. The final three columns of the table compare SD modelling with two commonly used economic modelling approaches, namely GEM and EconX.

Table 3.3 Sustainability problems and relevance for system dynamics modelling

Problem Methodological answer Ranking of SD Ranking of Ranking of (ex 6) GEM (ex 6) EconX (ex 6) Human-Nature Interdisciplinary approach 1= 5= 5= interactions Uncertainties Uncertainty management 4= 4= 3 Temporal externalities Long range view 1= 3 4 Spatial externalities Local global perspective 4= 4= 6 Social externalities Stakeholders participation 2 5= 4 Overall 2 4 5 Source: Based on Boulanger and Brechet, 2005

Boulanger and Brechet (2005) evaluate the six modelling tools based on the five methodological problems. The overall ranking of system dynamics modelling for each of the sustainability criteria is given in Table 3.3. These results should be interpreted with caution since all evaluations contain an element of subjectivity. However, the authors found that system dynamics performed better than both GEM and econometrics in terms of interdisciplinary potential, long term perspective and participation, performed the same as GEM in terms of uncertainty management and local global perspective, while econometrics performed better than SD in terms of uncertainty management and slightly worse in local global perspective. Overall, SD approaches score much better than GEM and EconX applications.

Min Kang and Jae (2005) use a slightly different categorisation of models. Four quadrants are proposed with static dynamic on one axis and feedback versus laundry relationship3 on the other. GEMs are dynamic yet laundry, regression and correlation is static while causal mapping is static with feedback. Computer simulation on the other hand is dynamic with causal feedback.

3.2.3 Evaluating alternatives

Although system dynamics (SD) modelling and traditional neoclassical (NC) theory share much in common (Crookes 2012), the modelling approach of NC and SD differ quite significantly. In that

3 Term used by authors referring to a linear relationship excluding any possible feedback relationships.

regard SD modelling and NC modelling represent different paradigms. The first major difference is that NC modelling is based only on historical data which limits the scope of information that can be used. Jay W Forrester, the founder of system dynamics, puts it like this: “To use only numerical data and to exclude information from the mental and written data bases means that one loses most of the available information about structure and governing policies” (Forrester 2003:9).

A second difference between NC economic modelling and SD modelling is that the former focuses on predicting accurate trends in economic variables. SD modelling, on the other hand, utilises a different kind of forecasting: “A system dynamics model should be used to forecast how the nature of the behaviour of a system would be altered by consistently following an alternative policy. Such a forecast of the ongoing effect of an enduring policy change can be done and can lead to improved systems” (Forrester 2003:10).

A third difference between NC economic modelling and SD modelling is that the former emphasises on models based on the price mechanism. SD economic models, on the other hand, allow for a much broader scope of behaviour: “Inventories, backlogs, and delivery delays are the primary short-term balancing forces. Prices then change as a result of over or under supply of product.” (Forrester 2003:10). Furthermore, SD models focus on disequilibrium conditions, and do not contain references to supply and demand curves as these relate to equilibrium conditions.

The important point that needs to be made here is that such observations do not advocate the universal rejection of neoclassical models. However, given the temporal and spatial scale of the current research problem, the complexity of the system (many and heterogeneous entities, and complex interactions between the entities) suggests a relative strength for SD modelling in this instance.

3.3 Applications in environmental-economic systems

Allison and Hobbs (2004) used a combination of resilience theory and system dynamics to identify causal relations and macro level system structure in the Western Australia (WA) agricultural region. The authors use five ecological social and economic variables that characterise the system. The ecological variable is the area of productive land, which in WA is classified into six major types: primary native vegetation, cropland, pastureland, commercial plantations, secondary native vegetation or regrowth, and unproductive land. The conceptual model of landuse change for this system is depicted in Figure 3.1:

Figure 3.1 A conceptual model of landuse change in Western Australia Source: Allison and Hobbs, 2004.

Other social and economic variables include the number of agricultural establishments, farmer age (discussed qualitatively), farmer terms of trade, and the wheat yield. Data used range from 1900 to 2000. As for the NM model discussed previously, not only historical dataa was consulted, but reference modes were interpreted with reference to the adaptive cycle metaphor and the Kontratiev cycles. Evidence from WA agricultural regiion suggested synchronisation of these adaptive cycles with the Kontratiev wave.

The dynamics of landuse change indicaated a progression from primary native vegetation to a productive agricultural system. Resilience steemmed from functional reinforcement within scales, and adaptive capacity between the social, economic and the ecological systems. Macroeconomic conditions were the primary drivers of the dynamics of the agricultural reggion, but were also influencced by institutional aspects such as government policy and markets.

In order to assess current usage of SD modelling for economic-environmental modelling, Science Direct was searched to obtain a suite of applications in water, agriculture, biodiversity and restorattion that were deemed particularly relevant to this study4. Since the current project is focussed on South Africa, a specific focus of these searches was on applications that model a particular geographical area. Therefore, models that model the ‘world’ or are generic studies that use hypothetical data or data that are not linked to a geographical area in the publication were excluded. In this way 35 case studies of relevance to this study were identified from peer-reviewed academic papers, mostly published between 2000 and 2010 (see Table 3.4 for full list of case studies; note that reports are excluded and therefore this list cannot be considered exhaustive but only indicative). Approximately half of these case studies (46 percent) utilised the VensimTM modelling software, although this result is mainly due to the nature of the search process that was used since this author was specifically interested in Vensim applications. A notable feature of recent system dynamics applications is the wide geographical distribution of these case studies. Furthermore, a

4 A number of economics case studies were alsoo identified. However, many of these weere generic models that addressed a theoretical question and were not for a specific geographical area.

number of economic-environmental modelling applications have already taken place in South Africa (Higgins et al. 1997; Jogo and Hassan 2010).

Table 3.4 System dynamics case studies referenced for geographical map, sorted by year

Author Year Category Location Software 1 Jogo and Hassan 2010 Water Limpopo, RSA STELLA 2 Jeong 2009 Other economic- S.Korea Vensim environmental 3 Khan 2009 Water Yellow River Basin, China Vensim 4 Jin et al. 2009 Other economic- Chongqing, China Vensim environmental 5 Guimarães et al. 2009 Agriculture Brazil Vensim 6 Nobre et al. 2009 Agriculture Zhejiang, China Powersim 7 Bendor 2009 Wetland restoration Chicago, USA STELLA 8 Videira et al. 2009 Water Portugal Not specified 9 Wang 2008 Other economic- Dalian, China Vensim environmental 10 Liu et al. 2008 Restoration, water Sichuan, China Vensim 11 Arquitt and 2008 Restoration Thailand Vensim Johnstone 12 Chung et al. 2008 Water Arizona, USA Powersim 13 Zhang et al. 2008 Water Tianjin, China Dynamo 14 Ford et al. 2007 Other economic- Washington Vensim environmental 15 Meerganz von 2007 Water Canary Islands, Spain Vensim Medeazza and Moreau 16 Liu et al. 2007 Water Guangdong, China Vensim 17 Ulli-Beer et al. 2007 Other economic- Switzerland Vensim environmental 18 Yeh et al. 2006 Agriculture Taiwan, China Vensim 19 Chen et al. 2005 Water Taiwan, China Vensim 20 Min Kang and Jae 2005 Other economic- S.Korea Vensim environmental 21 Shi and Gill 2005 Agriculture China STELLA 22 Guerrin 2004 Agriculture Reunion Vensim 23 Patterson et al. 2004 Agriculture, other Domimica STELLA environmental 24 Stave 2003 Water Las Vegas, Nevada Vensim 25 Güneralp and 2003 Water Turkey Not Barlas specified 26 Santos and 2003 Agriculture Portugal STELLA Cabral 27 Saysel et al. 2002 Agriculture Turkey STELLA 28 Guo et al. 2001 Water Yunan, China Not specified 29 Saysel and Barlas 2001 Water Turkey STELLA 30 Portela and 2001 Other environmental Brazil STELLA Rademacher 31 Vezjak et al. 1998 Water Slovenia STELLA 32 Alam et al. 1997 Agriculture Bangladesh Not specified

Author Year Category Location Software 33 Turpie et al. 1997 Other environmental Western Cape, RSA STELLA 34 Bockstael et al. 1995 Water, agriculture Maryland, USA STELLA 35 Bala et al. 1988 Agriculture Copenhagen Dynamo The cases of systems dynamics applications have grown rapidly in recent years. For example, in terms of the Vensim models reported on earlier, 88 percent of the studies have taken place in the past 5 years. The main reason for the rapid growth in this modelling technique in recent times is largely due to more powerful personal computers that have enabled complex simulations to be undertaken faster and by a wider audience, and also the advent of modelling software that has simplified the model building process (e.g. Voinov 2008). The results from this assessment suggest that system dynamics models are increasingly gaining recognition as a modelling technique relevant for studying integrated economic-environmental problems.

3.4 Steps in the modelling process

3.4.1 Comparison between different approaches

A number of authors have proposed explicit steps to be followed in the process of building systems dynamics models (e.g. Grant et al. 1997, Ford 1999, Forrester 2000). Table 3.5 compares these steps with the stages proposed in the systems engineering field. It is evident from this that there is considerable uniformity on the process to be followed in building a system dynamics model. This chapter uses a combination of steps based on the SD stages of NORBE (2004), Ford (1999), Sterman (2000) and Grant et al. (1997).

Table 3.5 Comparison between Systems Engineering (SE) and System Dynamics (SD) methodologies

SE stages SD stages SD stages SD stages (Norbe 2004) (Ford 1999) (Sterman 2000) (Grant et al. 1997) 1. Problem definition 1. Get acquainted with the 1. Problem articulation 1. Conceptual model system formulation 2. Goal setting 2. Be specific about the 2. Formulation of dynamic problem dynamic hypothesis 3. Systems synthesis 3. Construct the stock flow 3. Formulation of a 2. Quantitative model diagram simulation model specification 4. Draw the causal loop diagram 5. Estimate the parameter values 4. Systems analysis 6. Compare model to 4. Testing 3. Model evaluation reference mode (model validation) 7. Conduct sensitivity analysis 5. System selection 8. Test the impact of 5. Policy design and 4. Model use policies evaluation

Model building is an iterative process (Ford 1999; Sterman 2000; Grant et al. 1997). A model is usually built up in stages characterised by increasing complexity (Ford 1999). Furthermore, the phases are not always pursued sequentially, and certain phases may be repeated more than once (Grant et al. 1997). The model building process is terminated when the model is capable of replicating the observed behaviour of the system (Ford 1999). An illustration of the model building process is given in Figure 3.2.

Figure 3.2 Iterative model building process Source: Adapted from Sterman, 2000.

Sterman (2000) proposes a number of activities associated with each of the modelling steps (Table 3.6).

Table 3.6 Steps in the model building process

Steps Activities 1. Problem statement Define the problem Identify the key variables Identify the time horizon Historical behaviour of key variables 2. Formulation of dynamic Initial hypothesis: what are the current theories of behaviour? hypothesis Develop dynamic hypothesis Develop maps of causal structure, based on initial hypothesis, key variables, reference modes using causal loop diagrams, stock flow maps and other facilitation tools 3. Formulation of a simulation Specification of structure model Estimation of parameters Consistency tests 4. Testing Comparison to reference modes Robustness under extreme conditions Sensitivity analysis on uncertainty initial conditions 5. Policy design and evaluation Develop scenarios Identify new policies that may be designed in the real world Identify the effects of policies (What if analysis) Conduct sensitivity analysis Explore interactions between policies Source: Sterman, 2000

3.4.2 Conceptual model formulation

Model conceptualisation is an important stage in the model building process. Many systems problems are complex and multidisciplinary, with different actors having different perspectives on the problem. Using systems dynamics modelling provides a means of building consensus and focuses attention on the key driving forces that affect the problem at hand (Winch 1993).

The practitioner needs to obtain a sufficient understanding of the system in order to develop a realistic portrayal of the problem at hand. The modeller needs to manage the risk of an extreme ‘insiders’ perspective, or at the other end of the scale, too much of an ‘outsiders’ perspective (Mass 1986). Too much of an insider’s perspective can hinder the development of a new view of the system that causes decision makers to have a better understanding of policy choices. Avoiding an ‘outsiders’ perspective means that decision making processes need to be captured in ways that are familiar, meaningful, and representative of the real world system.

According to Forrester (2000), there are three stages to conceptual model formulation. Firstly, background knowledge on how feedback loops operate and guide examination of the problem. Secondly, the information gathering stage takes place which might include interviews with key personnel. “These interviews are extensive and penetrating. There may be several sessions with each of many individuals. The discussions range widely from normal operations, to what is done in various kinds of crises, what is in the self-interest of the individual…what would be done in hypothetical situations that may have never been experienced, and what actions are being taken to solve the serious problem facing the company” (Forrester 2000:13). Thirdly, a case study approach is adopted where the problems are described in words. A descriptive case study model is developed. “Such a descriptive case-study type of model is equivalent to a high-order nonlinear difference equation” (Forrester 2000:14)

3.4.3 Development of a dynamic hypothesis

Kirchner (1984) distinguishes between two modelling approaches: the advocacy strategy and the strategy of multiple hypotheses. For the advocacy strategy, the strongest case is made for a particular model or theory. It is characterised by a search for confirmatory evidence. The dominant theory is only modified when doing so will result in making it more defensible. The method of multiple hypotheses, on the other hand, is a process of searching among a set of credible alternative hypotheses. The search is for ways of disproving a hypothesis, and this narrows down the range of options. A third approach, proposed by Sterman (2000), involves selecting a hypothesis as a working hypothesis that explains the dynamics characterising the problem based on the feedback and stock and flow structure of the system. The approach is dynamic because it characterises the dynamics of the stocks and flow structure of the system. It is also provisional because it is subject to revision and abandonment as a better understanding of the system and real world processes are obtained. All of these approaches have their strengths and weaknesses. At the time of publication (1984) Kirchner observed that the most common technique employed by systems dynamics modellers was the advocacy strategy. There is some evidence that this still holds today. For example, Lane (2000b) argues that a ‘dynamic hypothesis’ embodies the concept that “a certain causal structure explains a certain dynamic behaviour”. He goes on to argue that “model building tests this hypothesis using

rigorous formulation” (p.4). However, the modeller needs to be able to adapt the modelling hypothesis if evidence from the model or from outside contradicts the hypothesis.

3.4.4 Quantitative model specification

This involves translating the model into a simulation model. “Such a model allows the computer to act out the roles of each decision point … and feed the results to other connected decision points to become the basis for the next round of decisions” (Forrester 2000:14). Quantification often leads to a better understanding of the structure and dynamics of a problem (Sterman 2002) that may result in reformulation of the problem. It may also assist in the identification of qualitative variables that may or may not be included in the model5.

According to Ford (1999), there are three stages to quantitative model specification: 1) Construct the stock flow diagram; 2) Draw the causal loop diagram 3) Estimate the parameter values. A number of tests may also be applied during the model building stage (Forrester and Senge 1980).

3.4.4.1 Stock flow diagram

The stock flow diagram is the basic building block of the system dynamics model, and represents the key linkages in the system, with a functional relationship behind each of these linkages. The stock flow diagram is elaborated on further in Section 3.5.2.

3.4.4.2 Causal loop diagram

While the stock flow diagram represents the linkages between each of the endogenous parameters in the model, the causal loop diagram (CLD) represents feedback in the system. There is no clear rule as to whether or not stock flow diagrams or CLDs are constructed first or either whether CLDs are necessary to the model building process (e.g. Ford 1999). Two types of feedback loops are possible: reinforcing feedback loops and balancing loops. Figure 3.3 illustrates the types of feedbacks possible.

5 Increasingly, qualitative data such as customer satisfaction and product quality are included in systems dynamics models and quantified using a number of social science techniques (see for example Luna-Reyes and Andersen 2003)

young fishes + + mature + births R B deaths + fishes - -

average lifetime

Figure 3.3 Simple causal loop diagram to illustrated balancing and reinforcing loops Key: births, deaths = variables + positive polarity – negative polarity R = reinforcing (positive) loop B = balancing (negative) loop. The parallel lines in the figure indicate a delay

In Figure 3.3, the positive sign on the causal link (e.g. between births and young fishes) suggests that an increase in births results in an increase in young fishes, and conversely that a decrease in births results in a decrease of young fishes. The negative polarity (e.g. on the link between deaths and mature fishes) indicates that an increase in deaths reduces the number of mature fishes. A reinforcing (positive loop) indicates that, on the whole, the cycle of births to young fishes to mature fishes increases the mature fish population in perpetuity (in the absence of other influences), while the balancing (negative) feedback loop reduces mature fishes. The direction of the arrow around the loop identifier moves in the same direction of the loop it identifies. In the above example both loops are clockwise.

3.4.4.3 Estimating the parameter values

Estimating the parameter values in the model involves quantifying the linkages in the model. According to Graham (1980), parameter estimating techniques fall into two categories: cases where data are at the level of aggregation of the model and those where data are below the level of aggregation of the model. Examples of the former are the most straightforward. If the model variable is quality of housing, then actual data exists on housing quality. The method of estimation is through the utilisation of model equations. However, one of the strengths of the system dynamics modelling approach is through utilisation of techniques to estimate parameter values in data poor environments. In the case where data are below the level of the model, data on housing quality do not exist and data below the level of the variable need to be used, for example average age of household units. There are two ways of obtaining this information (Graham 1980).

The first way of obtaining unknown sub-aggregate data may be termed descriptive methods. One (time consuming) way is to survey a number of households in the area of the study and determine the average age of the houses. Another way is expert input (asking a housing specialist or property agent the average age of houses). A third way is through visual observation. Are the houses characteristic of a particular period? Another option is to read histories of a neighbourhood to determine the average period that houses were built. Another method is for the modeller to utilise his or her own experiences in estimating the average age of house. Finally, a general estimate may be obtained by picking extreme ranges for the variables (outside which the value will not fall) and then picking a value in between. This final technique seldom needs to be used.

The second category of methods for obtaining unknown parameter estimates may be termed analytical approaches. Two approaches fall within this category: table functions and ad hoc approaches. Table functions use lookup tables to relate known parameter values to unknown parameter values. These lookup tables may either be based on actual data, on reference modes or known functional relationships. Ad hoc-approaches use known data to estimate values for the unknown parameters. The data may be available or collected by the modeller. A functional relationship is then developed to relate the known data to the unknown data. One way in which this may be done is through the technique of optimisation.

3.4.5 Model validation

The process of validation involves subjecting the model to a series of tests. It is important to realise that no model can be verified or validated because all models are wrong since they represent limited, simplified representations of the real world (Sterman 2000). Models, including system dynamics models, fail the extreme ‘Popperian’ approach of refutability and falsification (Sterman 2000). However, many of these models may then be ‘saved’ by adopting an auxiliary hypothesis.

This is true in other areas of science as well. Sterman (2000) gives the example of Galileo dropping balls of different weights from the leaning tower of Pisa6. The new hypothesis of balls of different weights hitting the ground at the same time disproved the previous hypothesis that balls of different weights descend to the ground at different rates, a true ‘Popperian’ approach to science. However, the ‘new’ hypothesis was later disproved by more accurate measuring, but redeemed by invoking the auxiliary hypothesis that air friction has an influence on speed of descent.

The important aspect of systems dynamics modelling, and models in general, is to recognise that their benefits lie in the ability to assist with decision-making (Sterman 2000). The objective for the modeller is to make the best model available for the purpose at hand in spite of its inevitable limitations. The important aspect in validation is to highlight the limitations of the model to decision- makers, so that it will not be misused and so that the model can be improved.

Model validation involves subjecting the model to a series of tests in order to ascertain whether or not it is a realistic portrayal of the system it is trying to model. Coyle and Exelby (2000) define validation as “the process by which we establish sufficient confidence in a model to be prepared to use it for some particular purpose” (p.28). A range of authors describe the main model tests used in validate a System Dynamics model (Forrester and Senge 1980; Richardson and Pugh 1981; Barlas 1989; Barlas 1996; Sterman 2000; Schwaninger and Groesser 2009, Hill 2010). Given the range of tests available (Schwaninger and Groesser 2009 alone discuss 24 tests), an important validation question is: which tests are mandatory and which tests are for reference? In this chapter we adopt a simple rule of thumb. We compare the list of tests proposed by five leading system dynamics practitioners, and the greater the number of practitioners utilising a particular test, the greater the weight given to the test results. This method is not without flaws since certain newer tests may have emerged that are as important as the older tests, but nonetheless provides an initial guideline. Ultimately, all models are wrong (Sterman 2000) and the aim of these tests is to build confidence in

6 Whether or not this was in fact how Galileo conducted his experiment is subject to speculation.

the usefulness of the model for answering the research questions. The range of tests proposed by these authors is summarised in Table 3.7. We employed all tests recommended by at least three of the five system dynamics practitioners (the cells in bold), with the first four tests (structure verification, parameter verification, dimensional consistency and boundary adequacy) given the most precedence. The highlighted cells indicate the tests actually implemented. Descriptions of each test (below) are drawn from these authors.

Table 3.7 Summary of validation tests conducted by different authors

Forrester and Richardson and Sterman, 2000 Schwaninger Hill, 2010 and No. Senge, 1980 Pugh, 1981 and Groesser, pers. comm. used 2009 (ex 5) Structure Face Validity Structure Structure Structure 5/5 verification assessment examination verification (implied) Parameter Parameter Parameter Parameter Parameter 5/5 verification values assessment examination verification (implied) Dimensional Dimensional Dimensional Dimensional Dimensional 5/5 consistency consistency consistency consistency consistency Boundary Boundary Boundary Boundary Boundary 5/5 adequacy adequacy adequacy adequacy adequacy Extreme Extreme Extreme Extreme 4/5 conditions conditions conditions condition Surprise Surprise Surprise Surprise 4/5 behaviour behaviour behaviour behaviour Behaviour Parameter Sensitivity Behaviour 4/5 sensitivity sensitivity analysis sensitivity Behaviour Behaviour Behaviour 3/5 reproduction reproduction reproduction Behaviour Behaviour Behaviour 3/5 anomaly anomaly anomaly Family member Family member Family member 3/5 Integration error Integration error Integration 3/5 error Behaviour Behaviour 2/5 prediction anticipation System System 2/5 improvement improvement Extreme policy 1/5 Mass balance 1/5 check Loop dominance 1/5 Turing test 1/5 Notes: Hill, pers. comm. is based notes made at an informal discussion with Andrew Hill of Ventana Systems, UK (date: 17th September, 2010) on methodology employed during client consultations and may therefore not represent an exhaustive list of validation methods employed either by himself or Ventana Systems UK. Schwaninger and Groesser (2009) also discuss a number of context related tests that are not discussed by any of the other authors

3.4.6 Model Use

The objective of the modelling effort is to develop a model that can ultimately be used for policy design and evaluation. According to Grant et al. (1997), there are four steps within the model use phase of system development: 1) Develop and execute the experimental design for the simulations; 2) analyse and interpret the simulation results; 3) Examine additional management policies or situations; 4) Communication of the simulation results.

In the model development phases, a runtime version of the baseline model is developed for use in policy simulations. A number of structural and dimensional consistency tests are executed to ensure that the model conforms to specifications.

In the model execution phase, a number of policy simulations or evaluations are undertaken depending on the objectives of the study. At the same time, the model is tested against key assumptions and extreme values (‘shocks’). Sometimes it may be necessary to return to earlier phases of the model development process, such as reformulation of the problem statement or the dynamic hypothesis. Once a working version of the model is available that passes a satisfactory range of tests the next step is to analyse and interpret key results.

In the third step, the model is examined with the purpose of suggesting ways that the system performance may be improved by making recommendations for future interventions in the system. The limitations of the model and recommendations for future refinements are also elaborated on.

Finally, model results are communicated. In a research setting, this usually means publication in a scientific journal (Grant et al. 1997). In a management framework, this implies communication of model results to those managers whose policy decisions impact on the area of study.

3.4.7 Tests for building confidence in models

Different tests are applicable at different stages in the model building process. These include tests of model structure, model behaviour and testing a model’s policy implications (Forrester and Senge 1980). The different tests are illustrated in Figure 3.4. There is no single test that serves to validate a model (Forrester and Senge 1980). However, confidence in a model gradually increases as the model passes more tests, and more closely resembles empirical reality.

Formulation of Policy design and STAGE Problem definition Testing simulation model evaluation

Structure verification Behaviour reproduction System improvement TESTS No tests identified Extreme conditions Behaviour prediction Changed behaviour Parameter verification Behaviour anomaly prediction Boundary adequacy Family member Boundary adequacy Dimensional consistency Surprise behaviour Policy sensitivity Extreme policy Boundary adequacy Behaviour sensitivity

Figure 3.4 Tests appropriate at different stages of the model building process Source: Based on Forrester and Senge, 1980

Tests of model structure include testing whether or not the model replicates the actual structure of the real world system it seeks to replicate, whether the dimensions (units) are consistent, the model must be tested under extreme conditions (that might never occur in reality). Finally, model structure is tested to ascertain whether the level of aggregation is appropriate and that the model contains all relevant parameters and feedbacks. The latter is done by comparing the model to a relevant hypothesis based on real world considerations.

Tests of model behaviour include how well the model behaviour replicates the behaviour of the real world system. If anomalies do occur between behaviour and reality, these need to be traced back to model assumptions. Sometimes it is also necessary to test whether or not the model contains features of a class (family) of models of which the specific model makes up a subset. Furthermore, it is necessary to understand whether surprise behaviour represents an anomaly or whether this is a feature of the model that has previously gone unnoticed. Testing the model resilience under extreme policy conditions related to rate equations (e.g. employment and personal savings flows fall to zero). Sensitivity of model parameters to changes in assumptions is also important, as are tests whether model structure are adequate to address the issues the model was designed for.

Tests of policy implications involve testing whether the policy implications resulting from a change in a parameter value in the model replicate the policy response that a real world system would predict. Tests in this category include the system improvement test, the changed behaviour prediction test and policy sensitivity test.

Not all tests are utilised all of the time when building systems dynamics models (Forrester and Senge 1980). Tests of model structure are most frequently used (e.g. Barlas (1996), followed by tests of model behaviour, and finally tests of policy implications (Forrester and Senge 1980).

3.5 Model features

3.5.1 Modelling software utilised

A number of different software packages are may be used to model system dynamics problems. For a recent comparison between these different modelling packages see Voinov (2008). This study uses the VensimTM Simulation Environment developed by Ventana ® Systems, Inc. (Ventana Systems 2007). Vensim is an interactive modelling environment that allows the development, calibration, simulation and optimisation of continuous simulation problems (Eberlein and Peterson 1992).

Some of the basic features of the modelling software include (Voinov 2008): stock flow modelling; finding the best match between data and model behaviour; optimisation using the efficient Powell hill climbing algorithm (Powell 1964), Kalman filter (Kalman 1960), Monte Carlo analysis, Causal Tracing® that highlights a selected variable in a tree structure that shows the variables that cause it, and many other features.

The software comes in a number of different versions, ranging from a model viewer, to a free PLE (Personal Learning Edition) and versions with more advanced functionality (PLE plus, Professional and DSS). The software version used in this study is the Vensim® DSS for WindowsTM Version 5.9e.

3.5.2 Basic building blocks

The main feature of system dynamics models are stocks (also known as reservoirs, levels or state variables), flows (rates or processes) and auxiliary variables (or converters) (Deaton and Winebrake 2000, Ford 1999, Güneralp and Barlas 2003). The stocks (represented by rectangles) are key variables in the sense that they represent accumulations in the system (Table 3.8). Flow variables (illustrated by valves) represent change in the system, activities which fill or drain the stocks. “Since a rate is really a mathematical first derivative of a variable… (system dynamics modelling) is usually equivalent to a set of first-order differential equations” (Robertshaw et al. 1978).

Table 3.8 Basic entities in a stock flow diagram Name Vensim symbol Description

Stock or level A component where a process accumulates Stock (for example water in a bath)

Rate or flow A measure of the rate at which a stock flow accumulates (e.g. volume of water per minute)

Connector Connects components and also indicates direction of causality Auxiliary variables (sometimes represented as circles) are either constants, or calculated from stocks, constants and other auxiliary variables. In Vensim, three types of conventions are often used

(Table 3.9). Variables in lower case represent standard auxiliary variables, a constant is given in capital letters, and a shadow variable in grey shading with brackets around it.

Table 3.9 Three types of variables used in the model Auxiliary variable variable name A component that interacts with other components (usually through a mathematical relationship)

Constant CONSTANT Distinguished from an auxiliary variable through the use of capitalisation

Shadow variable A variable that links with another variable/model in a different view

An illustration of how these different entities relate to each other in a stock flow diagram is given in Figure 3.5. This diagram show a simple stock flow diagram for fishes growing to maturity with a delay function associated with fisheries catch.

EFFECT OF FISH CARRYING DENSITY ON DEATH RATE CAPACITY fish density

death rate

Young Mature Fishes Fishes births maturation deaths TIME FOR FISHES TO MATURE BIRTH RATE perceived mature catch TIME TO CATCH fishes TIME TO COUNT FISHES FISHES

TARGET LIVE MATURE FISHES extra fishes

Figure 3.5 A System Dynamics model for fish population dynamics Source: Millennium Institute, 2010.

3.5.3 Dynamic behaviour

Most systems exhibit multiple feedback loops (Robertshaw et al. 1978) which allows for non-linear and counterintuitive behaviour and dynamics (Forrester 1971). Behaviour over time is characterised both by equilibrium and non-equilibrium dynamics (Deaton and Winebrake 2000, Radzicki 1997). Type of dynamic behaviour includes linear growth (straight line and positive slope) or decay (straight line and negative slope), exponential growth or decay (where the rate of growth or decline increases rapidly over time), logistic growth (also known as s-shaped growth, initially increasing and then declines as the population approaches carrying capacity), goal seeking behaviour (which is similar to

exponential decay except that, instead of seeking a goal of zero a non-zero goal is strived towards), oscillation and overshoot and collapse (which occurs when the system overshoots and the carrying capacity is permanently damaged, for example through overgrazing, resulting in the system being able to support a lower population than it would have initially). Examples of the different system features are given in Figure 3.6. For example, three types of oscillating behaviour are shown: damped, sustained and exploding.

Linear Growth Linear Decay

100 60

85 50

70 40

55 30

40 20 0 5 10 15 20 25 0 5 10 15 20 25 Time (Month) Time (Month) Current Value : Current Current Value : Current

Exponential growth Exponential decay

8,000 60

6,000 45

4,000 30

2,000 15

0 0 0 5 10 15 20 25 0 5 10 15 20 25 Time (Month) Time (Month) Current Value : Current Current Value : Current

S-Shaped growth S-Shaped decay

1,000 1,000

750 750

500 500

250 250

0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 0 10 20 30 40 50 60 70 80 90 100 Time (Year) Time (Month) Current Value : Current Current Value : Current

Goal seeking behaviour A Goal seeking behaviour B

60 40

50 30

40 20

30 10

20 0 0 5 10 15 20 25 0 5 10 15 20 25 Time (Month) Time (Month) Current Value : Current Current Value : Current

Oscillating behaviour A Oscillating behaviour B

100 200

85 150

70 100

55 50

40 0 0 5 10 15 20 25 0 5 10 15 20 25 Time (Month) Time (Month) Current Value : Current Current Value : Current

Oscillating behaviour C Overshoot and collapse

40 1,000

20 750

0 500

-20 250

-40 0 0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50 Time (Month) Time (Year) Current Value : Current Current Value : Current

Figure 3.6 Examples of dynamic behaviour in systems models

3.6 Conclusion

SD modelling approaches can be used to deal effectively with complexity and dynamics in a system. SD models are complementary to the more specific and static scientific applications, such as ecological, economic and hydrological studies on a site-specific level.

SD modelling approaches are increasingly gaining recognition as a modelling technique relevant for studying integrated economic-environmental problems and contain specific generic features and logical modelling steps that ensure replicability. In fact, several software packages using SD modelling approaches have been produced already, ensuring internal consistency and increasing confidence in the results supporting decision-making processes.

CHAPTER 4 OUTCOMES OF THE SYSTEMS DYNAMIC MODELLING AND PORTFOLIO MAPPING PROCESS7

4.1 Introduction

Douglas Crookes, Martin de Wit and James Blignaut

The aim of this chapter is to present the results from the Regional Economic SysTem dynamics mOdel for the Restoration of Ecosystems and project Prioritisation (The RESTORE-P model). The research is founded on the testing of the following hypothesis (see also Chapter 1):

The restoration of natural capital improves water flow and water quality, land productivity, in some instances sequesters more carbon, and, in general, improves both the socio- economic value of the land in and the surroundings of the restoration site as well as the agricultural potential of the land.

Most governments, organisations and individuals interested in restoration are constrained by budgets. Which restoration projects should then be undertaken? The aim of this modelling exercise is to classify and prioritise restoration projects subject to financial resource constraints. The technique employed to do this is the risk analysis (RA) framework proposed by David Hertz (Hertz and Thomas 1983). A more recent discussion of the approach is given in Aven (2003). The risk analysis approach uses Monte Carlo simulation to assign a probability distribution to the output variable. The output from the risk analysis process is then used to inform a portfolio mapping (PM) exercise maps (Matheson et al. 1989; Matheson and Menke 1994; Cooper 2005; Wysocki 2009) used to select and prioritise restoration projects. The approach adopted here is novel in that a system dynamics (SD) model of the problem is first developed, and then used as part of the risk analysis process. This is the first known application of risk analysis (RA), SD and PM to an environmental restoration problem.

4.2 Literature review/background

4.2.1 System dynamics and restoration

RNC is acknowledged as a complex systems problem and there is no single simple answer or single discipline capable of addressing the problem in question. There is a need to adopt synthetic approaches which integrate the dynamic and complex ecological and socio-economic aspects of the

7 This chapter is largely based on Crookes (chapter 4, in progress). Modelling the ecological-economic impacts of RNC, with a special focus on water and agriculture, at eight sites in South Africa. Ph.D thesis. Department of Economics. Stellenbosch University.

issues related to the restoration. System dynamics modelling is the ideal tool to understand and model these aspects.

System dynamics models are used for a wide range of economic and environmental applications. A review of a range of economic, hydrological and ecological journals indicates that although system dynamics modelling has been used to model restoration activities, its application has been limited as generally it is difficult to identify exactly what is meant by a restoration project. A number of restoration activities have focussed on wetland or watershed problems. Bendor (2009) developed a STELLA model for wetland restoration in Chicago. Liu et al. (2008) modelled watershed restoration in Sichuan Province, China and Arquitt and Johnstone (2008) model mangrove restoration in Thailand. On the other hand, applications of the system dynamics modelling technique to water, agricultural and other environmental problems are widespread. In southern Africa, system dynamics modelling is increasingly gaining prominence. For example, Higgins et al. (1997) model the restoration of mountain fynbos ecosystems in the Western Cape, Jogo and Hassan (2010) model wetland management in the Limpopo river basin, and Fleming et al. (2007) model cholera health risk. As far as we are aware no study has considered the impacts of restoration across a wide range of habitats and ecosystems using a system dynamics modelling framework.

4.2.2 Product and process innovation

The economics of innovation provides an important foundation for the restoration projects, as these entail innovative processes with uncertain benefits that remain untested to a large extent. In particular restoration projects are examples of ‘eco-innovations’ that are characterised by product and process innovation (see e.g. Hellström 2007). In order to further develop these concepts, it is important to define several innovation ‘adjectives’ (Swann, 2009; Henderson and Clark 1990). Incremental innovation describes the progressive increase in a product or process which does not change the character of the product or process in a fundamental way. An example would be improvements in the features of a motor cycle to enable it to go faster or make less noise. Component innovation involves the changing the core design concepts of a technology without changing the linkages between the core concepts and the components. Using the existing analogy, a component innovation may involve developing a motor cycle that runs on biofuel rather than petrol. Architectural innovation involves a fundamental change the way that a product is assembled, while leaving the components unchanged. An example of this using the motor cycle analogy is for the manufacturers to develop a motorised go-cart using the engine and other components of the motor cycle. Finally, radical innovation involves fundamental improvements that alter the character of a product or process. The development of a new car that undermined the competitiveness of the existing car manufacturers would be an example of this.

It is possible to distinguish between different approaches to Ecosystem Goods and Services (EGS) based on the framework proposed by Henderson and Clark (1990). In the first (incremental) case, EGS is seen a product that is already accepted on the market. The focus is on ensuring improvements in the way that these products are delivered to the market. An example would be the building of dams to provide water, or other innovative approaches to supplying EGS. The second (component) approach argues that EGS is already a marketed good, but argues that the RNC represents a change in the core design concepts of supplying EGS. The third approach, architectural

innovation, argues that EGS is a fundamental new product on the market that has not yet been fully exploited, but which can be marketed without changing any of the underlying processes that produce it. In other words, existing methods of environmental supply are seen as adequate. Finally, radical innovation regards EGS as a product that has not been fully exploited to date, and could potentially undermine other existing products on the market.

The radical innovation approach would be regarded as the Schumpeterian process of creative destruction while Swann (2009) argues that the Adam Smith’s concept of innovation through division of labour results in incremental innovations. Hellström (2007) notes, that most innovation (including eco-innovation) takes place in the incremental mode but that there is a “concomitant need to understand what types of eco-innovation can also be labelled radical innovation with high sustainability potential, and what specifically characterizes these” (p.149). The kind of innovation that follows from the combination and reorganisation of existing but previously distinct knowledge competencies (attributed to Arthur Koestler) is what is required for radical innovations (Swann 2009).

4.3 Data and materials

4.3.1 Study description

The RESTORE-P model is a localised system dynamics model that investigates the impacts of restoring natural capital across eight independent and largely heterogonous case study sites throughout South Africa (Figure 4.1 – see also Chapter 1).

Figure 4.1 Geographical distribution of case studies

The sites represent a range of vegetation types, from Nama Karoo and Succulant Karoo, to Fynbos, to Savannah, Grassland and Forest (Table 4.1). The majority of the sites are in arid or semi- arid climatic zones, with mean annual precipitation of less and 700 mm per year. Most of the restoration takes place on private land, although some have mixed landuse while others are public or communal areas. Most sites have low connectivity of the origin (measured as a rehabilitated area of less than 50 km2). The extent of degradation also varies quite significantly across the sites, and although this is difficult to compare with any degree of objectivity, many sites were severely degraded. Most notable were those affected by mining activity (strip mining) and those affected by intensive Ostrich farming as well as the communal grazing areas.

Table 4.1 Landuse and ecosystem characteristics of the eight study sites

Site Name Vegetation Climatic Mean Annual Land Area Extent of type Zone Precipitation classification rehabilitated degradation (km2) S1 Namaqualand Succulent karoo Arid 160 Private 26 Severely degraded: Restoration of sand dunes in the succulent karoo following open-pit surface mining in the Namaqualand S2 Beaufort Nama karoo Arid 239 Public/Private 8 Degraded: Clearing of invasive West alien plants (notably prosopis) in the Nama karoo S3 Oudtshoorn Succulent karoo Arid 242 Private 1762 Severely degraded: Restoration following overgrazing by ostriches in the succulent karoo S4 Lephalale Savanna Semi-Arid 400 Private 9249 Degraded: Bush-thinning (and combating bush encroachment) in the bushveld/savannah S5 Agulhas Fynbos Semi-Arid 478 Public/Private 548 Degraded: Clearing of invasive alien plants in the fynbos ecosystem S6 Kromme Fynbos Semi-Arid 650 Private 46 Degraded: Clearing of invasive alien plants in the riparian ecosystem S7 Drakensberg Grassland Temperate 900 Communal 1 Severely degraded: Restoration of a communal grassland system following overgrazing near the Okhombe village S8 Sand Forest/Savanna Temperate 1275 Public/Private 32 Degraded: Removal of exotic plantation forestry

4.3.2 Data sources

Primary data is derived from a range of studies conducted at the individual case study sites by a number of the co-authors (Cloete, forthcoming, Crookes 2011a, 2011b, De Abreu 2011, Fourie et al. 2011, Gull 2012a, 2012b, Marx 2011, Mugido 2011, Mugido and Kleynhans 2011, Ndhlovu 2011, Nowell 2011, Pauw 2011, Rebelo 2012a, Vlok 2010a, 2010b). Most parameter values were obtained from these published literature sources, from unpublished data that accompanied this research, or through personal communications from a range of experts. In a few cases where literature estimates were not available, the system dynamics model was utilised to optimise input parameters in such a way that Net Present Values (NPVs) for a particular case study were maximised. For example, the model indicated that optimal restoration period was an initial high level of activity followed by a

maintenance period, or a long term period of restoration activity at relatively lower intensity. See Annexure A for a full list of sites, parameters and sources used in the system dynamics model, as well as the equations in the model.

4.4 Conceptual model

The RESTORE model evaluates the effects of restoration on all four forms of natural capital (see e.g. Aronson et al. 2007), namely renewable, non-renewable, replenishable and cultivated (Figure 4.2). Five main types of ecosystem disturbance affect the sites in various combinations.

STRIP UNSUST. UNSUST. BUSH IAPs DISTURBANCE MINING FARMING FORESTRY ENCROACHMENT (-)

SITES S1 S2 S3 S4 S5 S6 S7 S8

NATURAL CULTIVATED REPLENISHABLE RENEWABLE NON-RENEWABLE CAPITAL

PRODUCTION AGRICULTURE WATER ECOSYSTEM BIOMASS MINING (+)

Game, agronomy, Lumber, GOODS Heavy horticulture, livestock, Yield, quality Soil carbon fuelwood, AND minerals SERVICES apiculture, wild products electricity

ECONOMICS Restoration costs Restoration Benefits method

Capital depreciation, labour, equipment, bond Value of improvements in different capital classes refinancing costs

Figure 4.2 Conceptual model: Five forms of environmental disturbances are present among the eight study sites covering four forms of natural capital diminishing the level and quality of the ecosystem goods and services emanating from these sites. This change in ecosystem goods and services largely affect agricultural productivity, water provision, carbon sequestration and climate amelioration and mineral deposits. The interaction between the cost and the benefit of the selected restoration method leads to an evaluation of the economic returns of the investment, while such an action replenishes the natural capital stock at the eight sites.

Five main types of ecosystem disturbance affect the sites in various combinations. The spread of Invasive Alien Plants (IAPs), and the closely related plantation forestry (which also utilises exotic

species) are two dominant disturbances that affect a number of different sites. At the Agulhas and Kromme sites, Black Wattle (Acacia mearnsii) and Port Jackson (Acacia saligna) and Blue gums (Eucalyptus spp.) are major invading plants, as well as Pines (Pinus spp.). At Beaufort West, Mesquite (Prosopis spp) is the dominant invader. At the Sand river site, a number of plantations exist, as well as invasives that include those mentioned for the Agulhas and Kromme sites, as well as Bugweed (Solanum mauritianum), Lantana (Lantana camara), Guava tree (Psidium guajava), Brambles (Rubus cuneifolius), Mauritius thorn (Caesalpinia decapetala) and Prickly pear (Opuntia stricta) (Le Maitre et al. 2002). Alien species are associated with a number of negative impacts (De Wit et al. 2001), inter alia: reduction of surface streamflow, loss of biodiversity, increases in fire hazard, increases in erosion and the destabilisation of river banks. Aliens do however also provide a number of benefits, including significantly in the context of this study firewood and other biomass energy (Agulhas and Beaufort West), timber and pulp (Sand) and also nitrogen fixation (Beaufort West).

The Namaqualand case study is the only example of natural capital that is not renewable. The mine produces 125,000 tons of zircon, 25,000 tons of rutile, 200,000 tons of pig iron and 200,000 tons of titania slag, but this comes at a great cost to the environment (Pauw, 2011). The strip mining process removes natural vegetation and top soil is removed to a depth of 50 mm. The subsoil is then removed and mixed with sea water, where the mined product (heavy minerals) are then separated from the fine particles (slimes) and tailings (oversized particles). Although much of top soil and tailings are utilised in the rehabilitation process, the natural vegetation is not able to fully recover, partly due to salination, partly leaching and partly wind erosion. As a result, a societal deadweight loss occurs, mainly affecting agricultural production in the area.

Although an improvement of farm production is a major beneficiary of restored landscapes, in a number of the case studies farming actually causes the degradation. In Oudtshoorn, Ostrich farming is a major cause of rangeland degradation (De Abreu, 2011). Ostriches strip off leaves and uproot plants, and furthermore trample the soil that lead to compaction and the removal of the biological crust and also creates pathways that lead to erosion. In the Drakensberg, poor management practises such as overgrazing and unseasonal burning of rangeland areas may affect species richness, the removal of vegetation cover, water infiltration rates, and significantly increase erosion (Marx, 2011). Trampling of ground, especially after rain events, creates pathways that eventually lead to erosion and the creation of dongas. In the Sand river study, it is not agriculture per se, but the canal system used to provide water for irrigated crops that has caused a number of problems. The canal system is significantly damaged and requires major repair work to fix leakages and also broken weirs (Crookes, 2011b). The consequence is that firstly not all irrigated crops receive water, and secondly, water users downstream of the agricultural zone do not receive the ecological reserve (the minimum water flow required to sustain ecological functioning). At the Kromme river site, farming production has resulted in eradication of large areas of the indigenous palmiet wetlands (since this area has the most fertile soil) (Rebelo, 2012b). Human intervention has also reduced the frequency of fires in the Lephalale study site, resulting in increased bush encroachment and/or bush thickening. For restoration to be effective in these areas, it is crucial that rehabilitation proceeds in conjunction with improved management practices.

The model is characterised by a number of reinforcing and balancing feedback loops. The causal loop diagram for each of the eight case study sites is roughly the same and comparable with the

generic version presented in Figure 4.3. The first feedback loop indicates the relationship between an ecosystem disturbance and the natural vegetation. An increase in unsustainable use (in the absence of restoration) creates a negative feedback loop that increases ecosystem disturbance and results in a reduction in natural capital. This indicates that the more abundant the natural capital the greater the risk of unsustainable use. This is an example of negative (or balancing) feedback loop, and has the tendency to produce stable, equilibrium or goal seeking behaviour over time. In other words, the system will stabilise but not necessarily at a sustainable level.

ecosystem goods and services +

+ economic species value R composition +

+ natural capital+

+ -

unsustainable use B restoration R - + ecosystem ecosystem disturbance - functionality

Figure 4.3 Causal loop diagram. This diagram shows the balancing (B) and reinforcing (R) loops linking economic value to natural capital stock, restoration and other variables.

Positive (or reinforcing) feedback loops have the effect of generating increasing or amplifying model behaviour. There are two reinforcing effects in the model: one an ecosystem feedback loop and the other an economic feedback loop. An increase in unsustainable use increases the ecosystem disturbance, which reduces ecosystem functionality that increases the need for restoration. An increase of restoration, on the other hand, increases natural capital. In the economic feedback loop, an increase in natural vegetation increases species composition which increases ecosystem goods and services which increases economic value. It is important to emphasise that the link between economic value and natural vegetation is a policy link and is not explicitly modelled in the system dynamics model. In other words, a positive economic value is used as a justification for PES which in turn has the potential to further increase the area under natural vegetation.

4.5 Research method

In the first instance, a system dynamics model was developed and used to maximise the net present value of the each of the eight case studies. The model was developed in Vensim DSS 5.9e (Ventana Systems, 2007). This platform provides an interactive modelling environment for answering policy related questions. The platform also enables the identification of key structural features in the model, as well as conducting the Monte Carlo simulations used for sensitivity analysis (model validation) as well as risk analysis.

The interrogation of the model indicates that price, in particular the water value but also sensitivity in grazing and crop input prices, is the main driver in the system and indicated the major area of uncertainty in the model. We develop this further by conducting multivariate simulation on the main decision variables in the model, in order to forecast the distribution of the payoff variable, in this case NPV. A number of different distributions are possible for the payoff variable. These include the Normal, Poisson, Uniform and Triangular distributions. Usually, the uniform distribution is utilised if no additional information apart from the ranges in key variables is known (Van Groenendaal and Kleijnen, 1997). In this case, input parameters were described using the uniform distribution, with the degree of variation reflecting the uncertainty of the parameter. Future refinements of the model should focus on obtaining a better understanding of underlying distribution functions characterising the model.

Parameter values for all simulations were standardised to ensure comparability across study sites. Since input prices could potentially range across any positive value up to and including the baseline, minimum values for the price function assumed -100% of the baseline value (i.e. zero), with maximum values equal to the baseline. Monte Carlo simulations were conducted for an ensemble of 200 realisations, for crop values driven price, the value of water based a spectrum of water values, and the economic value of grazing capacity improvements (resulting in an increase of potential stocking rates) as a result of restoration. Uncertainties in the input parameters lead to uncertainties in the output parameters. Summary statistics for the output variables are given in Table 4.2. In most cases uncertainties in the output parameters are less than uncertainties in the input parameters, since the standard deviation is less than the mean (or the coefficient of variation (CV) is less than 1).

Table 4.2 Monte Carlo summary statistics for output variable (NPV, t=50, 2060)

Water Crop Grazing mean StDev CV mean StDev CV mean StDev CV (Rm) (Rm) (Rm) (Rm) (Rm) (Rm) Namaqualand -99.205 0.520 na Beaufort West 1.344 1.295 0.964 2.938 0.045 0.015 Oudtshoorn 30.745 9.050 0.294 Lephalale -2435.787 1326.260 na Agulhas 176.808 116.423 0.658 375.476 0.003 0.000 Kromme (a) 105.379 107.656 1.022 285.261 1.721 0.006 275.943 7.701 0.028 Kromme (na) 22.112 68.260 3.087 Drakensberg 0.222 1.154 5.200 2.185 0.003 0.001 Sand 348.348 240.607 0.691 426.167 79.997 0.188

From the output of the Monte Carlo simulations, it is also possible to compute the probability of success of a project, measured as the number of realisations (out of 200) that contain a positive NPV.

4.6 Results

4.6.1 Value of the payoff variable

Spatial area of restoration has a significant impact on overall restoration impact. Results indicate a fairly substantial positive contribution of restoration across most sites (Table 4.3 and Figure 4.4). The NPVs ranges from –R96 million to almost R480mil after 50 years, once all ecosystem benefits are taken into consideration (cultivated, replenishable and renewable).

Table 4.3 Summary of PVs for the different restoration sites (R million)

Area PVs of restoration (R million) Total NPV Ha Cost Benefits R million Cultivated Replenishable Renewable Namaqua 2,619 -93.4 -3.1 0.0 0.7 -95.8 Beaufort West 781 -2.3 0.2 3.9 1.3 3.0 Oudtshoorn 176,216 0.0 30.8 0.0 15.4 46.2 Lephalale 924,920 -5,523.1 5,090.1 0.0 910.2 477.3 Agulhas 54,755 -12.1 0.0 419.4 -8.6 398.6 Kromme-agric 4,640 -98.9 34.0 358.6 0.0 293.7 Kromme- no agr 4,640 -98.9 0.0 240.7 0.0 141.8 Drakensberg 90 -2.0 0.0 3.9 0.3 2.2 Sand 3,216 -24.3 368.7 304.3 -35.9 612.8

Eliminating the spatial component of the data indicates a relatively low value associated with more arid areas, and a higher value for more temperate areas. This result, however, ignores risk and uncertainty.

Figure 4.4 Cumulative NPV (R/hectare) for each of the different study sites Notes: sites arranged in decreasing order of ariidity. B1 indicates the present value (PVV) of restoration costs plus PV cultivated benefits; B2 indicates B1 plus replenishable natural capital PV, and B3 is B2 plus renewable PV (i.e. total NPV). (S = Sand river, Dg = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Kr = Kromme; Ou = Oudtshoorn)

4.6.2 System dynamics model

The system dynamics model was firstly used to understand the dynamics of the system. For example, in the Agulhas case study the fire regime played an important role in informing the impact of restorration and the potential economic return (Figure 4.5).

area mature fynbos npv agulhas 80,000 400 M

298.5 M 40,000 197 M Hectare Rand

95.5 M 0 2010 2025 2040 2055 -6 M 2010 2015 2020 2025203020352040 2045 2050 20552060 Time (Year) Time (Year)

Figure 4.5 Illustration of the dynamics of the systems model

The system dynamics model was also used to test key tipping points in the model. For example, sensitivity of the profitability of the biomass electricity plant was not significant for the Beaufort West net present value calculations (Figure 4.6a). However the system in general was sensitive to fluctuations in the water value, as indicated by the figure from the Sand river systems model (Figure 4.6b)

Current Current 50% 75% 95% 100% 50% 75% 95% 100% npv beaufort west npv sand 6 M 2 B

4.5 M 1.498 B

3 M 997 M

1.5 M 495.5 M

0 -6 M 2010 2016 2023 2029 2035 2041 2048 2054 2060 2010 2016 2023 2029 2035 2041 2048 2054 2060 a) Time (Year) b) Time (Year)

Figure 4.6 Monte Carlo simulations for: a. changes in profitability of biomass electricity plant b. changes in the water value Finally the system model was used to inform the risk analysis process. The volatility of the payoff function indicated the degree of uncertainty associated with the input shares. For example, for the Drakensberg site (Figure 4.7a), the effect of water price uncertainty produced a relatively smaller ‘plume’ compared with water price uncertainty at the Beaufort West site (Figure 4.7b). The figure also indicates the probability that a project is successful (measured by the area of the ‘plume’ that is above zero). The volatility of the output variable (NPV) to input uncertainty is measured by the standard deviation and coefficient of variation of the output variable. All three of these variables are used in the risk analysis component of the study.

Current Current 50% 75% 95% 100% 50% 75% 95% 100% npv drakensberg npv beaufort west 4 M 4 M

2 M 2.75 M

0 1.5 M

-2 M 250,000

-4 M -1 M 2010 2016 2023 2029 2035 2041 2048 2054 2060 2010 2016 2023 2029 2035 2041 2048 2054 2060 a) Time (Year) b) Time (Year)

Figure 4.7 Monte Carlo simulations for price sensitivity of water for a. Drakensberg b. Beaufort West Note: Parameter values for all simulations were standardised to ensure comparability across study sites. Since input prices could potentially range across any positive value up to and including the baseline, minimum values for the price function assumed -100% of the baseline value (i.e. zero), with maximum values equal to the baseline.

The width of the confidence plume is an indicator of risk. The wider the plume is, the higher the degree of uncertainty. A visual assessment of a diagrammatical representation is, however, potentially flawed as the scales on the Y-axis differ. An algebraic calculation is required to ascertain true volatility.

4.6.3 Portfolio mapping

Portfolio map classifications: The final stage in the risk analysis process, namely portfolio mapping, provides a means of selecting and • Oysters: high risk projects with prioritising between restoration projects. Three portfolio uncertain merits, maps are given, highlighting different aspects of portfolio • Pearls: projects with high risk. likelihood of success, • 4.6.3.1 Project costs Bread and Butter: essential projects that enterprises cannot do without, and The standard and most commonly used portfolio map is • White elephant: projects which the risk-reward bubble plot (Figure 4.8), with the size of are preferable to avoid. the bubble indicating resources committed to it. This (Note: This mapping classification is provides the means of comparing projects by considering derived from, among others, Cooper a range of factors (reward or payoff, probability of et al. 1997 & Cooper 2005.) success, and cost). It should be noted that these projects are not independent of each other, so the total resource cost will not add up to the budget. Furthermore, although some projects indicate a negative NPV, this is only because the project costs are compared with one ecosystem good at a time, rather than the entire range of EGS that are assessed for the project as a whole.

Results indicate that water projects are the ‘pearl’ projects, with high expected success likelihoods and high payoffs. Grazing and crop projects are mostly the bread and butter projects. There is one white elephant, the Namaqualand mining project, with large resources committed to it.

It should however be noted that this excludes the value of the benefits from mining, which would affect the feasibility of restoration. Mining benefits are omitted from the analysis since mineral extraction is not a renewable resource and therefore not sustainable under a strong sustainability perspective. Lephalale (grazing) is a potential oyster, with untested and therefore uncertain long term benefits from restoration. Fairly low levels of resources are committed to this activity.

1.2

0.8

0.6 Grazing Water -100000 -500000.4 0 50000 100000 150000 Crop

0.2

-0.2

Figure 4.8 Portfolio map for different ecosystem services: bubble size indicates resources committed to it (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn)

In spite of the usefulness of this portfolio map, a limitation is that the risks inherent in each project outcome are not highlighted. As a result we need to consider the second portfolio map.

4.6.3.2 Standard deviation

The second portfolio map is plotted against the same two axes, but instead of the size of the bubbles representing costs of restoration, the standard deviation of each project is included (Figure 4.9). The standard deviation indicates the degree of volatility of the data, and shows that, for the most part, the higher the potential reward the higher the risk. The projects with the most volatility are the water projects, as well as the irrigated agriculture scenario in the sand project. Most projects with low NPV (the so called ‘bread and butter’ projects), exhibit very low project volatility.

1.2

0.8

0.6 Grazing Water -100000 -50000 0.4 0 50000 100000 150000 Crop

0.2

-0.2

Figure 4.9 Portfolio map for different ecosystem services: bubble size indicates standard deviations (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW =

Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn)

4.6.3.3 Coefficient of variation

The final portfolio map gives the coefficient of variation (CV) as bubble size, but because CVs cannot be calculated for negative means, white elephants are not shown (Figure 4.10). CVs are appropriate when the project means show a wide range of dispersion. The results are somewhat different from the standard deviation plots, and suggest that the Drakensberg water project, and the Kromme water project (no agriculture scenario) are perhaps better classified as oysters rather than pearls, given the high degree of volatility.

Prob. Bread and butter success 1.2 High Pearls Ou S Ka S 1 Ag Ka BW 0.8 Kna D Grazing 0.6 R/ha Water -100000 -50000 0.4 0 50000 100000 150000 Crop

0.2 Reward (NPV)

-0.2 White elephants Low Oysters

Figure 4.10 Portfolio map for different ecosystem services: bubble size indicates coefficient of variation (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn)

4.7 Discussion

The RA process finds that no individual measure of risk (success probability, SD, CV) is sufficient for selecting and classifying projects. A combination of measures provides an improved means of selection. This is then used to inform a portfolio mapping exercise, in order to classify and select restoration projects (Table 4.4). A summary of the classification of projects suggests that the projects with the highest potential payoffs (and therefore are pearl projects) are the water projects, in other words those projects where downstream water consumers benefit from the restoration project. Agulhas, Beaufort West, Kromme and Sand are all examples of this. These are potential examples of Koestlerian innovation, where a multidisciplinary approach may yield greater synergies.

Table 4.4 Summary of projects classified by type

Oyster Pearl Bread and butter White elephant Description High risk projects Projects with high Essential projects that Projects which are with uncertain likelihood of enterprises cannot do preferable to merits success without avoid Water projects Dg; Kr (na) Ag, BW, Kr (a), S) Crop projects S Ag, Kr(a) Grazing projects Lp Ou (passive only) BW, Dg, Kr (a) Nam

However, the results also indicate that within this subset of restoration studies water alone was insufficient to mitigate the risks of the project. Table 4.4 shows that those projects that include agriculture (in the mix) are subject to lower risk. Firstly, Kromme without agriculture is classified as oyster (in other words, more risky) compared with Kromme (with agriculture), which is classified as a

pearl. Secondly, in the Sand study, in the case where Sabie Sand Game Reserve only benefits water is a higher risk project compared with restoration where irrigated agriculture also benefits. Another restoration study which is too reliant on water for benefits is the Drakensberg study, which is also classified as an oyster. Communal agricultural benefits and carbon values are not sufficient to increase resilience in the system. Lephalale on the other hand, is too reliant on grazing, and the introduction of a biomass electricity plant could potentially mitigate that risk and even push the project into an oyster or bread and butter project.

The bread and butter projects are almost entirely crop or grazing projects, but these are only profitable if combined with either water or biomass projects. The bread and butter projects are examples of Smithian innovation, where the division of labour results in qualitative improvements in outcomes. These project benefits are essential to ensure the success of restoration activities. A diverse project portfolio requires both Koestlerian innovation and Smithian forms of innovation.

The analysis of projects using portfolio mapping suggest that this approach, coupled with risk analysis and system dynamics modelling, is able to provide a means of selecting and prioritising restoration projects. Caution should however be exercised in interpreting results. A positive NPV should not be interpreted as a license to exploit the natural environment. Another area where caution should be exercised in applying the results of this study is in the case of critical natural capital. In such cases restoration should proceed regardless of whether or not NPVs are positive (see e.g. Farley and Gaddis, 2007). This study does however indicate that economic criteria are not the only norm for determining restoration sites. Other criteria that consider ecological, technological and process risks are also important.

4.8 Conclusion

The focus of the project is on assessing the ecological, hydrological and economic impact of restoration across a range of contrasting sites in South Africa. Data on NPVs of individual restoration projects at eight sites were collected and analysed. The outcome indicates that estimating the NPV (the payoff) of a restoration project is not sufficient in order to determine the viability of a restoration project. The inherent risk associated with the project also needs to be assessed. In order to do this, an integrated RA/SD/PM model (the RESTORE-P model) to select and prioritise restoration projects is required. The analysis draws on a number of transdisciplinary perspectives, including the economics of innovation, inputs from ecology, hydrology and agriculture, risk analysis, system dynamics modelling and portfolio mapping. The results indicate that a plurality of methods that consider risk provides a far superior decision making framework for selecting and prioritising restoration projects.

CHAPTER 5 DISCUSSION OF THE RESULTS

Edited by James Blignaut

Here we provide seven lone-standing sections that deal with different higher-level issues and lessons that have been learnt during the course of this study. They are effectively “nuggets distilled after reflecting on five years’ of work”. The seven topics are:

1. The use of well-established existing economic evaluation tools within a restoration and systems dynamic context is likely to yield different outcomes than merely applying them in a conventional manner; 2. The development of a restoration prioritisation protocol following the insights gleaned from these eight studies; 3. Natural capital can, indeed does, play an important role as an insurance mechanism mitigating risk if managed well, the converse is also true; 4. Restoration can, and indeed does, offer relatively cheap alternative options for providing the Environmental Reserve – an increase in water flow following restoration is actually “new” water to the system, or it is the return of water that was stolen; 5. After a deliberate act (such as through mining) of environmental degradation, compulsory restoration is often required. Restoration, however, rarely returns the full complement of ecological functions and services. What should be done with the unmitigated damage cost following restoration? 6. How to bridge the major information divide between scientist and researchers? Here we discuss a process of stakeholder engagement. 7. Policy recommendations are made here.

5.1 Message 1: Economic evaluation tools and systems dynamic models are useful to inform decision-making on restoration

Martin de Wit and James Blignaut

5.1.1 Introduction

We used standard economic project evaluation methodologies in this study and applied them to projects that restore natural capital. Net present values were calculated, taking into account the costs and benefits of restoration projects over time. Costs estimates were based on actual annual operational costs of comparable restoration projects, and inclusive of the depreciation of capital and interest on financing if capital costs were involved. The economic benefits of restoration were calculated on the basis of ecosystem goods and services yielded before and after restoration, taking into account the site-specific time lags, and/or dynamics, for ecological benefits to become evident. Ecosystems goods and services that were quantified at different sites include additional grazing and browsing potential for livestock and game respectively, additional water flow and estimates of water yield, provisioning of biomass energy, recreational benefits from water flows as well as the ability to sequester carbon. The aesthetic benefits of intact landscapes to property owners were identified in specific areas, but not directly quantified in this study.

The economic evaluation is based on data generated by eight case studies countrywide, namely Namaqualand, Beaufort-West, Oudtshoorn, Lephalale, Agulhas Plain, Kromme River, Drakensberg and the Sand River. The study sites that were chosen reflect a broad range of biophysical parameters such as ecosystems (fynbos, desert, riparian, grassland and savanna), soil types and precipitation (a gradient between 160 mm/a and almost 1300 mm/a). The sites also varied with respect to socio- economic parameters such as the types of beneficiaries (e.g. farmers, rural and urban water users, tourists and recreationists), and the value they would be willing to pay for an increased flow of higher quality ecosystem goods and services.

The results are interesting, but not surprising and in line with published results on the costs and benefits of alien clearing and restoration conducted before in South Africa (De Wit et al. 2001; Turpie et al. 2008, Blignaut et al. 2010). In general, economic evaluation of these restoration case studies shows that the costs of clearing or restoration exceed the private benefits of restoration, and when social benefits are included they do exceed the costs, but only under certain conditions. Private benefits are mainly related to improved grazing for livestock and/or game, and biomass energy production in certain instances. The main reason for the lack of private benefits is the relatively high costs of restoration in the early phases of a restoration project, as well as the lagged, dispersed and scale-related nature of benefits. The value of improved grazing, water and carbon for the private land-owner is usually not high enough to justify restoration, simply because the economic activities on the land will not be able to offset the high input costs of clearing or restoration. The constrained economic returns from the land suppress the ability, and thus willingness, to pay for additional input of ecosystem goods and services. This may also reflect the

implicit subsidised nature of many farming activities in a country with large areas of marginally productive land, a shortage of water and a shortage of quality grazing in some areas. Similarly, the value of biomass energy to the private owner is also not high enough to pay for the clearing costs involved. It can be concluded for these cases that rural land-owners focussed on farming for their livelihoods are not able and willing to pay for restoration.

The next question is who else gains from restoration besides the rural land-owner whose livelihood depends on farming activities. The social benefits are mainly related to improved water flow and yield, and water-related services such as recreation but, in some cases, carbon or improved landscape scenery and views (see also Sections 5.3 and 5.4). In other words, restoring ecological infrastructure at one place would lead to an improved flow of ecosystem goods and services which have value elsewhere. The mechanisms of such ‘value transmission’ can be provisional (such as water flows and carbon sequestration), but could also be cultural or informational (such as improved recreation or aesthetics), or flow regulation (such as the provisioning of base flow and the ecological reserve).

An important potential tipping point in the economic viability of restoration for most of the cases studied is the value of water. In most cases the societal value for water is close to what is required for restoration projects to pass a standard economic cost-benefit test. In drought-prone areas with sizable populations such as Beaufort-West, the actual cost of water is already higher than the value required to make restoration viable. Restoring the Kromme river system also provides water at a levelised cost per m3 comparable to the lower-end supply augmentation options in the Nelson Mandela Metropole. In the Agulhas Plain societal water values required to render restoration projects viable are already on par to what users are paying in peri-urban areas. The point is that restoration interventions do generate benefits that are valued by society, and within certain conditions, provide sufficient incentive for payments for increased ecosystem services. The research demonstrates that the highest potential for payments for ecosystems services is among urban and peri-urban users of water.

One of the biggest barriers towards realising such societal values is the large upfront costs in restoration. This is a situation similar to all other infrastructure (manufactured or man-made) investments, a pertinent example being dams. Unit Reference Values, a measure of the costs for every one Rand of benefit (i.e. water) from dams often ranges from R4-R8/m3 (Ninham Shand 2009). This means that, on average, investments in dams in South Africa would not pass a standard benefit- cost test if the price of water is not set by at least these values – and it seldom is. Ecological restoration is also a water supply option and needs to be framed as an investment in ecological infrastructure that yields, among others, water services. In that sense, investment in ecological infrastructure, including restoration, needs to be evaluated according to economic evaluation rules of other infrastructural investments rather than those of private projects.

5.1.2 Economic tools need to be used with a sensitivity to dynamic realities

An aspect often overlooked in standard economic evaluation is the dynamics of time, also called temporal dynamics. This is not a weakness of the economic methodology in particular, but rather in

its limited application. It cannot be assumed that the future will be like the past, especially when working with complex and dynamic systems such as ecosystems. Ecosystems are complex and often display non-linear behaviour over time. Some of these complexities are known upfront, and some are not. Treating the passage of time explicitly as a variable that could influence results, and not just as static and exogenous, could certainly change the outcomes of economic evaluation. What is not economically viable in a certain year, may become viable in the next year or at a future point in time, depending on how the system behaves.

For example, in a context of increasing values for water due to increased supply-constraints of freshwater, rising costs of backstop technologies and/or increased demand, the economic viability of restoration projects may change rapidly. Restoration should therefore be seriously considered as a long-term water resource augmentation option, as well as providing other services. Not only changing value, but also changing restoration costs (influenced by e.g. scale, method, costs of labour, capital and technology) and/or changing insights into ecological functioning would lead to different economic evaluation outcomes on restoration interventions. All this shows that time matters, and a dynamic modelling of a restoration over time, according to best knowledge, generally leads to better insights into the temporal horizons over which restoration is expected to become economically viable.

5.1.3 Economic tools need to be used as if risk and uncertainty matters

As systems are increasingly accepted as complex, and temporal horizons are expanded, more uncertainty is introduced in economic evaluation. What may be shown to be economically viable on average may not be the case when greater weights are attributed to certain risks. Conversely, what might not be considered viable in the short-term under certain conditions, may become viable under different circumstances. Economic evaluation tools usually assume the future to be like the past, which is a very strong assumption and one subject to considerable uncertainty. In fact, because much can happen in complex systems – especially in open, living systems such as ecosystems – the potential risks need to be explicitly included in the evaluation.

It must be noted that fundamental uncertainties cannot be modelled, but this limitation applies to all evaluation on all types of projects. Therefore economic evaluation needs to be augmented with an approach that explicitly includes risks. In this systems modelling study we dealt with simulated risks through the use of a Monte Carlo analysis, producing an envelope of economic viability and highlighting the confidence levels at which restoration will or will not be viable. The results of such a dynamic risk analysis are important additions to the otherwise limited economic evaluation of restoration interventions. Some restoration projects are economically viable regardless of the risks, but some others are very sensitive to risk.

5.2 Message 2: Restoration potential assessment protocol

Karen Esler and David le Maitre

5.2.1 Introduction

South Africa is rapidly losing natural capital through unwise land-use planning and management practices. A number of national and local programmes have been developed to address various components of these losses, but these tend to focus on particular issues or symptoms because they are driven by particular sectors or stakeholders. For example, Working for Water targets invasive alien plants, Working for Wetlands focuses on wetland restoration, LandCare on emerging farmers and land degradation. These interventions are necessary but often overlook the broader context and fail to address factors that will determine their ultimate success. In many cases these initiatives are failing to build in resilience and sustainability and thus are not achieving their aims (e.g. van Wilgen et al. 2012). Institutional structures, mandates and jurisdictions also often drive a fragmented approach to problems that are, in reality, complex and require multi-institutional collaboration and multi- or trans-disciplinary approaches (Swilling 2010). The aim of this discussion is to sketch some of the complexities of developing a more broad-based approach and to offer some guidance on the kinds of approaches that could determine the outcomes of interventions aimed at restoring natural capital.

RNC seeks to rebuild ecosystem resilience, thereby replenishing natural capital stocks and improving the flow of goods and services to society. It recognises that these goods and services are interrelated and interdependent and that the ability to use them sustainability requires planning and management systems that cut across sectoral and institutional silos. Both resilience and sustainability have social, economic and ecological dimensions, all of which need to be addressed when guiding interventions aimed at restoring natural capital. As a result, prioritisation exercises for restoration cannot be developed as 'shopping lists'; in reality, a range of complexities need to be considered. We suggest that one route is to consider a systems-modelling-based approach – the power of which is demonstrated in Chapters 3 and 4.

5.2.2 Aspects and complexities that need to be considered

Institutional arrangements

The natural resource governance framework within South Africa is multifaceted and complex. Some national departments have full jurisdiction over certain natural resources (e.g. water) and others jointly govern natural resources (e.g. agriculture and environment) at both national and provincial levels. This means that funding and efforts tend to be pigeonholed when, in reality, cross-sectoral initiatives or responses would be more effective.

The country has developed and implemented a number of policies, planning frameworks and systems which are used by national, provincial or local government agencies. These generally use

terms like ‘integrated planning’ or ‘frameworks’ or ‘integrated development planning’ but, in practice, they tend to emphasise economic and social development and do not adequately address the ecological or environmental resource base (natural capital) that generates the ecosystem services on which such development depends.

Natural resource governance requires the involvement of a range of actors, interacting at various scales and with varying levels of power. The challenge is to find the right actors who can unlock potential by working together to overcome the other complexities mentioned below.

Equity and redress

One of the key aims of the post-1994 policies and government initiatives has been to address the inequities in access to, and benefits from, ecosystem services that are a legacy of the apartheid system. This issue is addressed in the section of the constitution on human rights and underpins, for example, the National Water Act, 36 of 1998 (some, for all, forever), and the emphasis on environmental justice in the National Environmental Management Act 107 of 1998 as well as being a key element of natural resource-related policies. The poor generally have the greatest dependence on natural capital and the services it generates, and under previous regimes were often crowded into relatively resource-poor areas. This injustice needs to be recognised when determining where programmes and projects aimed at natural capital restoration are located when they are being planned, initiated and assessed.

Natural capital and ecosystem services

The idea of natural capital was generated by the recognition that ecological systems contain species and communities of organisms which interact with each other and with the climate and physical environment and perform functions that sustain those ecosystems. These organisms, and the functions they sustain, yield services which also benefit people, including soil formation and stabilisation, nutrient and water cycling, water purification, production of food and medicines and providing places for recreation. Some of these services are localised i.e. the services and their benefits are available in particular location. This is typical of production services which yield foods and medicines. The ultimate products can be transported elsewhere, but their sources are localised. Other services can yield both local and non-local benefits. The regulation of air quality (e.g. the quantity of carbon dioxide) is a regional benefit while water flow regulation can be of benefit to people downstream, including in other countries. When guiding interventions aimed at restoring natural capital, we need to take cognisance of the multiple temporal and spatial scales within which these components are generated.

Ecosystem services generally also come as bundles of services and the nature of the benefits derived from these bundles can vary depending on their location. For example, a restoration project in the headwaters or a river far inland will have many downstream beneficiaries but on a short coastal river, there will be fewer beneficiaries. This spatial dimension is currently not well explored but can play a critical role in determining the mix of the services and the outcomes of restoration projects and so needs to be considered in targeting such projects.

Limited resources

There are insufficient resources to address all of the factors that are causing the loss of natural capital across the country, so the available resources needed to be targeted towards factors and towards areas where restoration will result in the greatest gains and sustainable solutions. To borrow a term from the conservation domain such areas could be termed restoration “hot spots”. Decision makers at various levels of government need guidance on how to identify such “hot spots” which should address key components or aspects of those decisions related to the governance, beneficiary and environmental domains (Figure 5.1). The aspects that we address at this level mainly relate to institutions and bodies in the public sector, but the private sector and civil society (including non-government organisations) are also key role players and need to be brought into the process and into such guidelines. The private sector, whether from the agricultural, mining or business sectors, operate well where markets for the benefits derived from ecosystem services exist, or can be created, and are, therefore, appropriate implementers for the RNC in some areas and situations. In other cases rural communities can be a more appropriate target.

Governance aspects: • Willingness of land owners / users to participate • Institutional arrangements • Government stimulation in the absence of markets • Policy and legal goals and imperatives • Sustainability / permanence • Equity and redress Environmental aspects: • Degree of aridness and fragility, • State of land: use and/or cover • Potential for state transitions and their reversability • Spatial and temporal dynamics RNC • Landscape connectivity or fragmentation • Geology (groundwater, production potential) • Conservation importance • Rate and or susceptibility to degradation (incl dam siltation) Potential markets/beneficiary aspects: • Link / proximity to users / beneficiaries • Quantified benefits / alternative users • Public / private cost and benefits • Equity and redress • Landscape position in relation to the kinds of services being addressed • Spatial scale?

Figure 5.1 The three dimensions and underlying factors that need to be considered when evaluating the potential for a programme or project aimed at restoring natural capital.

Variability and uncertainty

The world we live in is inherently variable, and uncertainty about the nature and consequences of such variability, especially in environments which are already experiencing climate change, is a factor that has to be taken into account. The tools that are typically used for assessing where, when and how to undertake projects or interventions, often treat the world as static and unvarying. For example, widely-used economic tools for project selection such as cost-benefit analyses assume that the state of the system being assessed does not vary significantly in ways which will affect the outcome. We argue elsewhere (Section 5.1) that this is not so and that modelling approaches which link economic, ecological and social systems models are needed to take temporal variation into account. For assessments that address large-spatial scales we need to be able to take both spatial and temporal variation and their interactions into account. Systems modelling, which links ecological, economic and social components, offers the potential to address both spatial and temporal variability, but the complexities this involves are not well understood and require further research.

5.2.3 Adaptive management as an approach

Adaptive management is an approach to resource management which recognises that we do not: (a) have perfect knowledge, (b) there is always uncertainty, and (c) complex systems generate surprises, and seeks to balance this against the need for clearly defined and achievable goals. It establishes a clear framework for guiding resource planning and management programmes and for engaging with participants and stakeholders in the establishment of the goals, the planning and the implementation. The broad-based approach advocated by adaptive management enables all three dimensions of the situation to be addressed. Guidelines for the selection, planning and implementation of restoration projects emphasise the importance of establishing clearly defined and achievable goals and targets. They also emphasise the importance of uncertainty and of natural variability in determining the outcomes of such restoration projects so there is close alignment between the thinking behind adaptive management and restoration ecology.

5.2.4 Towards guidelines

There are no universal guidelines or decision models that can be followed when identifying and prioritising RNC projects. Prioritization efforts in the past have largely drawn upon normative frameworks (which require an understanding of social, ecological and economic norms) but we argue that the complexity of the trade-offs and interactions between these norms necessitates a new approach. There are good reasons for including the wide variations in the mixes of ecosystem services that could be improved and the wide range of the needs of the potential beneficiaries and their ability to pay. Any such framework would need to take into account the ecological, social and economic risks as well as the potential returns on those investments. It would also need to address three main components (Figure 5.1): (a) issues of governance focused on how do we make this work?; (b) the environmental factors that can determine the options and the outcomes; and (c) who will benefit from this and how well matched are the benefits and their needs?

Based on experience in other contexts such as conservation planning (Cowling et al. 2008), it is critical that the process should be as participative as possible which, inevitably, means that every assessment will be customised and tailored to a greater or lesser extent and a single decision tree or fixed hierarchy of criteria would not be effective. Nevertheless it is important to provide some structure and guidance to the process to reduce the risks of making the wrong decisions.

Science should not have a special role or a privileged position but it can contribute to a better process. One way is to introduce better modelling approaches that incorporate the complexities mentioned above while still providing structure. We believe that systems modelling provides one such approach. These models have the potential to generate new, sometimes unexpected insights and understanding to the decision-making process.

The first step in generating the knowledge for systems modelling is to provide a non- threatening space in which inter- or transdisciplinary teams can interact to generate the wide variety of insights needed to generate interventions aimed at restoring natural capital. Team management must ensure that the process of generating and structuring the knowledge and interpreting the outputs is well-structured and managed for it to succeed. Above all, we need people and reciprocity to get the teams working – the willing seller, willing buyer scenario. Understanding who the key actors are, how they relate to each other and where the power-base lies is critical in finding the key actors who will unlock the potential for restoration success and overcome the challenges listed above.

Dealing with (environmental, economic and social) risk, variability and uncertainty are key challenges facing restoration prioritization programmes. In this study on the restoration of natural capital we succeeded in incorporating various layers of complexity including variation in services over time and space. However, we did not explicitly include the actor domain, although this can be achieved using a range of social research tools such as interviews and workshops, and then to integrate these into the systems modelling approach (Crookes 2012). In the end, it is the actors who make the decisions about resource allocation, and their voices must be incorporated into the process. The model described in Section 5.1 is a start towards capturing the spatial dimension to economic assessments in the same way that systems models are able to capture the temporal (varying time) dimension.

The items and issues listed in each of the components (Figure 5.1) should not be regarded as a check list where all conditions must be met before evaluating the potential for a programme or project aimed at restoring natural capital. Rather, where sets of factors coincide, these should be seen as indicators of opportunity. The trick is in recognising these opportunities and unlocking the potential.

Finally the old adage that nothing succeeds like success is highly relevant. Success stories have the ability to motivate others. On its web page, The Society for Ecological Restoration (http://www.ser.org/) has global examples of “before” vs. “after” restoration projects – we need to build on these examples to highlight our own, uniquely South African success stories.

5.3 Message 3: Restoration is an investment in natural capital and an insurance premium against risk

Steve Mitchell, Alanna Rebelo and James Blignaut

5.3.1 Introduction

The insurance industry at large is based on the underlying principles that:

1. not all people contributing to an insurance fund will be adversely affected simultaneously by an undesired yet plausible event, and 2. that it is wise/better to make provision in advance to either hedge and/or minimise the shock or negative impacts should an undesired event occur.

The combination of these two principles implies that it is prudent to diversify and use financial capital to act as a buffer against possible losses in manufactured capital (e.g. houses, cars), or cultivated capital (e.g. crops), or even human capital (e.g. the loss of the ability to earn and income or even death). Keeping capital intact is the underlying philosophy of modern economic development. The notion of using current streams of financial flows within the context of maintaining a desired level of future welfare is therefore well-entrenched within the general practice of investing in some form of insurance. An insurance policy is therefore a buffer against undesired yet plausible future adverse events. Within the context of global change, which includes but is not limited to climate change, the mass migration of humans and other species, pressures added by population growth and the need to improve the quality of life (particularly of the poor) the world has become increasingly complex and dynamic. This is exacerbated by the rapid development of telecommunication and ensuing mass-fragmentation of both information as well as social systems. Good examples are since modern communication (and the advances made in social network systems such as Facebook and Twitter) allows for a high degree of individualisation and customisation that was not possible before. Similarly, bio-physical systems are becoming fragmented due to rapid land transformation for a variety of purposes – the larger landscape-wide natural mosaic is lost among a web of individual land uses, i.e. intensive farming of various kinds next to one another, urban and rural settlements and conservation areas.

The ongoing fragmentation of both social and bio-physical systems increases the risk of system failure becoming a reality. In many respects it is now merely a question of “when is the next natural disaster”, and “when is it going to hit my village”.

In addition to having a financial insurance in place, there is an urgent need to put nature-based insurance policies in place to hedge against adverse effects on natural capital. The similarity between financial insurance and ecosystem-based approaches has been recognised (Van Oosterzee et al. 2012). Ecosystem services themselves, however, have no recognised hedge mechanism. In a method similar to that proposed by van Oosterzee et al. (2012) to hedge against risk in a REDD (reduction in emissions through degradation and deforestation) programme, here we propose a

mechanism through which the delivery of ecosystem goods and services on which we depend may be hedged against failure or malfunctioning of natural systems as a result of degradation and a loss in quality. We suggest that restoring functionality of degraded ecosystems will offer a bio-physical system-wide insurance against the effects of adverse events. Ecological restoration activity is therefore likely to safeguard against the loss of essential services, as well as against the loss in manufactured, cultivated and human capital. To illustrate this we will elaborate using the restoration of a riparian (riverine) ecosystem as one example.

The ecosystem goods and services most widely associated with a riparian ecosystem is the supply or provision of water for consumptive use. This is because our entire social and economic structure is completely dependent on water availability. Often the management of riparian zones focuses on the delivery of this service at a high assurance of supply, for example through the construction of dams or impoundments. Riparian ecosystems, however, deliver a wide range of other ecosystem goods and services (Box 5.1). Some of the other important services are listed in Table 5.1 together with the benefits that society derives from them.

A riparian system in good condition will deliver each of these services effectively, and at no financial charge to society. Maintaining the overall condition of the riparian system will ensure that the resilience is high, insuring against future losses of property and life. It is of the utmost importance to recognise that healthy riparian systems offer much more than just water supply. The converse is, however, is also true. A badly managed and/or degraded riparian system seriously compromises the ability of the system to provide the water provisioning service. It would seem counter-intuitive to allow a riparian system to become degraded given that this will reduce the potential for both water provision and its insurance function.

Box 5.1 Recommendations • Water Banks: Intact ecosystems function as water banks, hedging the socio-economic system against the deleterious effects of the extreme climatic events which are expected as part of global change. These need to be conserved so that they will continue to provide these goods and services. • Restoration: Where ecosystems have become degraded, they need to be restored to enable them to deliver the full suite of ecosystem goods and services of which they are capable, so hedging society against extreme climatic events through services such as functional water banks and flood attenuation. • Prevention rather than cure: The premium paid to keep ecosystems in good condition is lower than that paid to restore those that have been damaged. It makes sense to take the long-term view in ecosystem management, conserving what we have, rather than looking for short-term gains which will reduce the stream of goods and services in the longer term. • Who pays? The cost of restoring ecosystems is substantial and often there is a lag period between the outlay of costs and the return on investment. These costs should not be borne entirely by the landowner, but by society, either through the government or by the end users benefitting from the increase in ecosystem functionality. • Education & behavioural change: There is a need for society to change the way ecosystems are viewed. This should lead to a change in the way that they are valued by society. Once this change is embedded, the degradation of ecosystems will be stemmed and society will benefit from the hedging capacity that they offer.

5.3.2 Background: Water banks

In a generally arid region, such as most of South and southern Africa, impoundments are essential to provide the required assurance of supply to support socio-economic activities. In the dry season when there is no or little rainfall, especially in times of prolonged drought, impoundments may fail to meet the water demand. During these times, additional capacity may be provided from water stored in the natural infrastructure which, if allowed to naturally replenish during wet periods, would be able to release the water in the dry season. This source of water storage is especially important in areas of high climatic variability where it may be used to mitigate the uncertainty and so hedge the risk of not providing the required assurance of supply. Water may be stored in various compartments, or water banks, of the natural infrastructure. We discuss three such water banks here, namely wetlands, aquifers and the soil profile.

Wetlands

Wetlands, particularly valley bottom wetlands and peat mires, such as those seen in the Kromme and Mfabeni (Rebelo, 2012b; Grundling, 2012), are important in regulating water supply. Wetlands slow the velocity of water moving through the landscape and allow water to infiltrate into aquifers, replenishing these supplies for dry periods. Wetlands with substantial peat beds beneath them (for example the wetlands in the Kromme River have peat beds measured to be over 5 m deep) also act as a sponges, absorbing water into the peat which also acts as a pure carbon filter, improving water quality. However the optimum delivery of these essential services can only be achieved by wetlands in good condition. Wetlands that have been damaged can be successfully restored, although the recovery of the whole suite of services may only be seen after many years and the cost of this restoration may be substantial. However these costs need not be borne by individual land owners but by the downstream beneficiaries / users of the improved resource.

Aquifers

Aquifers are another important compartment in which water may be stored. Aquifers are valuable as water is stored underground and is thus not lost to evaporation. When water is stored above ground in impoundments, substantial volumes are lost to evaporation. Evaporation measurements on Midmar Dam during the first 2 weeks of July 2007 found that it occasionally exceeded 0.3 mm per 30 minute period over midday. This amounts to 3 m3 ha-1 per 30 minute period (Jarmain et al., 2009). Aquifers are recharged naturally in systems that have not been excessively damaged, through infiltration of water during the wet season and more so during a series of wet years. However damage to the natural system, such as destruction of wetlands and overgrazing, which by reducing vegetation cover decreases water infiltration, reduces aquifer recharge. This means that insurance against the dry seasons and drought years is diminished. However, recharge can be recovered in damaged systems by restoring ecosystem functionality and the volume of water held in the aquifer can be supplemented through artificial recharge. Examples of ecosystem restoration may be found in the studies in Oudtshoorn (de Abreu 2012), Beaufort West (Ndhlovu 2012, Vlok 2012), the Agulhas Plain (Nowell 2012, Vlok 2012), and the Kromme River (Rebelo 2012b). However if aquifers become polluted then it will be necessary to take further steps to treat the water provided by these aquifers

which will be an additional cost. These findings highlight the importance of preventing damage to natural infrastructure where possible, with restoration of these systems being a last resort.

Soil profile

The soil profile is a shorter term water bank than aquifers, with the water stored in this compartment being available to support plant growth through the subsequent dry season. As with aquifers, the water is protected from exposure to the sun and wind, the two ‘bank robbers’, so reducing loss through evaporation from the surface (there will still be losses through plant transpiration). It is, however, important to keep the soil in good condition if it is to optimally perform this function. This means that the organic and particulate structure must be such that infiltration into the soil profile will occur, and that the water will remain there and not infiltrate too deep or too fast for plant roots to reach. Overgrazing, too-frequent burning and damage to rivers cause the development of headcuts in wetlands or dongas, which drain water from the soil profile. This has occurred in many cases in South Africa, including but not limited to the Drakensberg (Marx 2012, Crookes 2012), Oudtshoorn (de Abreu 2012) and the Kromme River (Rebelo 2012). The restoration of dongas is an important but extremely costly intervention in the overall prevention of soil erosion and protection of this “water bank”. This is currently addressed through both the Working for Wetlands and the Mondi Wetlands Projects (WfWet, No date; Braack et al., No date).

5.3.3 Analysis: Riparian zones’ insurance capabilities – and restoration as an adaptation strategy

In addition to acting as important water banks, riparian zones also fulfil a life-supporting insurance function as discussed above and play other roles (Table 5.1).

Table 5.1 List of riparian ecosystems services

Ecosystem goods & Contribution to welfare and insurance capacity Plausible restoration interventions when system Examples of restoration in South services of riparian zones has become degraded, leading to a loss in Africa ecosystem goods and services Provisioning Water supply Increased water supply improves economic Restore wetland functionality Kromme River Study (Rebelo and Gull, welfare and livelihoods through increasing Remove woody alien trees from catchments 2012) agricultural potential and health through Stop siltation of dams by restoring eroded areas in Sand River Catchment Study (Crookes, insuring security of access to water in times of catchments (rehabilitating headcuts and stopping 2012) drought. overgrazing) Agulhas Study (Nowell, 2012; Vlok, 2012) Beaufort West Study (Ndlovu, 2012; Vlok, 2012) Plant materials (food, This insures the sustainability and improvement Sustainable catchment management: Drakensberg Study (Marx, 2012; medicinal, ornamental, of economic welfare and livelihoods as well as • Stop overgrazing/burning or other unsustainable Crookes, 2012) fodder and construction) food security. Wood for construction can land use management Oudtshoorn Study (de Abreu, 2012) & Biodiversity. provide shelter. Food and medical provision can • Clear IAP’s Lephalale Study (Cloete, 2012) also improve health and social cohesion. • Rehabilitate dongas to stop drop in water tables Increased food and medicinal provision can empower people to help others, so insuring societal stability. Improved biodiversity increases ecosystem resilience and stability, ensuring people the opportunity for freedom of choice and action. Fish and other fauna – Ecosystems in good condition ensure improved Improvement of water quality. Stopping pollution – Kromme River Study (Rebelo and Gull, Inland and estuaries food security, economic welfare and livelihoods chemical: (pesticides/herbicides) and biological 2012) through the provision of access to nutritious (eutrophication) –sewage. food which has health and welfare benefits. Restoration of wetlands and natural habitat of The increased opportunities for ecotourism aquatic organisms (fish and their food) –this may provided by intact ecosystems expand the require clearing of IAP’s. economic opportunities available, so assuring Removing alien fish. improved livelihood opportunities.

Regulating Carbon sequestration This is a larger scale benefit (national and Rehabilitating wetlands (peat mires in particular) Kromme River Study (Rebelo and Gull, international), but is important in mitigating the which act as carbon sinks. 2012) predicted effects of global warming. A decrease Integrated catchment management: restore Baviaanskloof PES Study (Mander et al. in atmospheric carbon may have implications for degraded rangelands (those damaged by overgrazing 2010) agricultural production and bush encroachment. or too frequent burning). Mfabeni (Grundling, 2012) Flow regulation Flow regulation ensures that water is available in Rehabilitating wetlands and riparian buffer zones Kromme River Study (Rebelo and Gull, (baseflow) the dry season when there is no rainfall, insuring (wetlands act as a sponge, slowing the movement of 2012) supply at times of water scarcity. This gives water and allowing it to infiltrate and replenish Sand River Catchment Study (Crookes, communities and municipalities water security, ground water supplies in aquifers) 2012) which translates into various economic and Clearing IAP’s (which use large amounts of water – social benefits. This also improves social sometimes even from aquifers) relations and cohesion and Improves habitats.

Groundwater recharge Ground water recharge increases water Restore/rehabilitate wetlands and riparian zones Kromme River Study (Rebelo and Gull, availability in the dry season both through water Remove woody IAP’s in riparian zones 2012) in the aquifers and through increased baseflow. Artificial recharge of suitable aquifers to increase the Drakensberg Study (Marx, 2012; This insures water security for communities and volume of stored water. Crookes, 2012) municipalities and translates into various Oudtshoorn Study (de Abreu, 2012)) economic and social benefits. This also improves Exxaro Study (Pauw, 2012) social relations and cohesion. Flood attenuation Flood attenuation through intact ecosystems Rehabilitating wetlands and riparian buffer zones – Kromme River Study (Rebelo and Gull, provides security from flood damage, protecting reduce erosion and capture sediment. 2012) lives, homes, agriculture and infrastructure. It Clearing IAPs from riparian zones and wetlands. Esler et al., 2008. also ensures that a greater proportion of flood Legislation such as CARA provides the framework for water is absorbed by the catchment to protecting riparian buffer zones and wetlands. contribute to the baseflow at a later time. Soil stabilisation This hedges against decreased quality of life by Rehabilitating wetlands and riparian buffer zones Drakensberg Study (Marx, 2012; reducing particulates in the air. It also provides a (wetlands trap sediments) Crookes, 2012) hedge against dam siltation, prolonging the Clearing IAPs from riparian zones and wetlands Oudtshoorn Study (de Abreu, 2012)) useful life of dams. It improves water quality (cause instability of riparian systems and wetlands). Kromme River Study (Rebelo and Gull, through the decrease in turbidity, providing Integrated catchment management: restore 2012) health benefits and saving costs in water degraded rangelands (those damaged by overgrazing Exxaro Study (Pauw, 2012) treatment. or too frequent burning). Stabilizing soil, especially top soil, both retains (not restoration) Careful planning of future the natural seed bank and improves agricultural infrastructure (roads/railways, etc.)

productivity. It provides religious & cultural benefits (such as grave sites) by protecting important heritage sites from destruction by dongas erosion. Water quality Intact ecosystems improve water quality, Stop point pollution (fertilizers/pesticides/ sewage Kromme River Study (Rebelo and Gull, (pollution assimilation & insuring against poor health through the from waste water treatment plants or from 2012) waste dilution) provision of clean water, saving on the cost of communities without infrastructure to remove raw water treatment. sewage) Water quality helps maintain riparian Rehabilitate wetlands and riparian buffer zones and ecosystems in good health, insuring resilience of clear IAP’s (wetlands and riparian buffer zones act as the ecosystem and providing food, welfare and a sink for pollutants) ecotourism benefits.

Supporting Habitat Provides the template for the maintenance of Integrated catchment management: biodiversity, insuring resilience and increasing • Clear IAP’s the value of recreation and ecotourism. • Stop/rehabilitate over-grazing Genetic / species This increases resilience of species which Integrated catchment management: Drakensberg Study (Marx, 2012; diversity contribute to biodiversity and sustainability, • Clear IAP’s Crookes, 2012) hedging against the loss of future options as well • Stop/rehabilitate over-grazing Oudtshoorn Study (de Abreu, 2012)) as loss of ecosystem services such as medicinal, • Landscape design – planning for resilience Exxaro Study(Pauw, 2012) food and fibre production in the long term. Beaufort West Study (Ndlovu, 2012; Increased resilience improves food security, Vlok, 2012a) economic welfare and livelihoods. It also insures access to nutritious food which has health and welfare benefits in times of duress. Genetic and species diversity insures against the loss of opportunities for recreation and ecotourism as well as the cultural value of ecosystems.

Cultural Cultural (including Insures against the destabilisation of societies Integrated catchment management: Drakensberg Study (Marx, 2012; spiritual) through providing social cohesion, mutual Crookes, 2012) respect, and freedom of choice. It also provides health and psychological benefits.

Tourism (aesthetic) Intact ecosystems provide an environment which Integrated catchment management: insures health and psychological benefits for • Clear IAP’s tourists. • Stop/rehabilitate over-grazing/unsustainable management

It is evident that society at large is gleaning Destroying our innsurance policy? multiple benefits from riparian systems, with well- The case off the Kromme managed and functioning systems acting as buffers and insurance against the impacts of thhe adverse effects of extreme events – all of which are becoming increasingly frequent and more severe. In addition to the above, effectively closing an all-important loop, restorattion restores ecosystem functionnality. The restored functionality has to be viewed as an important adaptive intervention aimed at mminimising the effects of global change, increasing the adaptive capacity and so the resilience of society in the face of the predicted changes. Overgrazing, too-frequent burning of vegetation, destruction and removall of wetlands, damage to river structure (canalisattion / building of roads / 5.3.4 Conclusion railway lines), and alien invasion has nearly brought ecosystem good and service delivery in the Kromme River to a sstandstill. Over the years Riparian systems are able to provide valuable goods as damage to the catchment has increased, the and services to society, offering a hedge against the water quality has steadily decreased (both potential damage caused by extreme events. chemically and in terms of silt load), erosion has increased, flood damage has increased and However, the value of the hedge is dependent on the there has been a steady decline in riverflow, condition of the ecosystem, because an ecosystem in especially in the dry season, and thus a decline good condition is able to provide a richer suite of in yield. services than one that has been degraded. Degraded However restoration, in terms of clearing out the ecosystems can and should be restored, although the alien invasive trees and rehabilitating wetlands functionality of the post-restoration ecosystem may (using gabions and weeirs) has the potential to increase riverflow – especially during the dry depend on how badly it was degraded. Unnmitigated season, and therefore yield, it will also increase damages might still occur even where restoration water quality by decreasing turbidity, and lastly appears to meet its stated objectives (see Section 5.5). flood damage will decrease (Rebelo 2012). Restoration is more costly than conserving ecosystems See also Section 5.6. in good condition, so it makes sense to manage what we have responsibly with the view to the long-term, sustained benefits from the provision of the water banks.

The cost of ecosystem restoration projects may be substantial and there is usually a lag period between the restoration intervention and thhe realisation of the benefits. It is not realistic to expect the full costs to be borne by the land owner while benefits are shared by broader society. The costs of restooring the water bank capability to hedge society against the deleteriouus effects of extreme events should be carried by society, either tthrough government or through the beneficiaries of the interventions.

The value of the riparian system is morre than meets the eye. The eye of the beholder needs re- calibration and this requires the mutual interaction of authorities, landowners / users, researchers and civic society at large. From this interactiion, a new frame of reference will be developed, leading to a new appreciation and hence valuation of the resource in the future.

5.4 Message 4: Complying with the environmental reserve: at whose and at what cost?

Doug Crookes and James Blignaut

An important innovation in the 1998 National Water Act of South Africa (RSA 1998; Ashton et al. 2011) is the introduction of the notion of a Reserve. The Act defines the Reserve as follows:

Reserve means the quantity and quality of water required -

(a) to satisfy basic human needs by securing a basic water supply, as prescribed under the Water Services Act, 1997 (Act No. 108 of 1997), for people who are now or who will, in the reasonably near future, be -

(i) relying upon;

(ii) taking water from; or

(iii) being supplied from, the relevant water resource; and

(b) to protect aquatic ecosystems in order to secure ecologically sustainable development and use of the relevant water resource;

The Act states further, in the preamble to Part 3:

The Reserve, … consists of two parts – the basic human needs Reserve and the ecological Reserve. The basic human needs Reserve provides for the essential needs of individuals served by the water resource in question and includes water for drinking, for food preparation and for personal hygiene. The ecological Reserve relates to the water required to protect the aquatic ecosystems of the water resource. The Reserve refers to both the quantity and quality of the water in the resource, and will vary depending on the class of the resource. The Minister is required to determine the Reserve for all or part of any significant water resource. If a resource has not yet been classified, a preliminary determination of the Reserve may be made and later superseded by a new one. Once the Reserve is determined for a water resource it is binding in the same way as the class and the resource quality objectives.

The class of the river is determined through a very rigorous process (DWAF 2007), mostly considering hydro-ecological attributes. This process considers, inter alia: the instream flow requirement, the water level, the presence and concentration of particular substances in the water, the characteristics and quality of the water resource and the instream and riparian habitat, the characteristics and distribution of aquatic biota, and the regulation or prohibition of instream or land-based activities which may affect the quantity of water in or quality of the water resource. One of the interesting features of the Reserve determination is that once a river class has been decided upon, that designation provides management parameters within which and towards which the river has to be managed, as is stated in the last sentence of the preamble to Part 3 of the Act quoted above. While such a categorisation is not static, the process of Reserve determination and

categorisation can be repeated in future and the class can therefore be adjusted, but for the time that a specific class has been decided upon it remains statutory (DWAF 2007).

In many cases, however, it has proved to be extremely difficult for the government to comply with the set classifications and to meet the prescribed environmental flows. For example in the case of the Sand River, the river has been assigned an environmental flow allocation of 43mil m3 per annum, but the system is oversubscribed and as a result it is running at a deficit of between 9mil m3 and 22mil m3 per annum (Agterkamp 2009). Because of this deficit a number of water supply augmentation options have been proposed, including the completion of the Bushbuckridge Transfer Pipeline (BTP) from Inyaka dam, renovation of the canal system, and changes in the water licensing and operating rules (Pollard et al. 2011). It is nonetheless recognised that full assurance of supply of the environmental flows is unlikely. This uncertainty is exacerbated by the high seasonal nature of rainfall. In addition the obligations to meet domestic requirements in the catchment place water for human needs in competition with water for environmental needs in the minds of many. However flawed this conceptualisation is, it does prevail with the Inter Basin Transfer now being prioritised as a domestic supply option rather than one to meet ecosystem security.

The non-compliance and/or non-delivery of the environmental flows implies that either the river class, and with it its Reserve requirement, has to be set lower, or more water has to be ‘found in the system’ which might mean that water has to be taken away from certain existing uses (authorised or not) to provide for environmental flows. Reducing the river class has both social and ecological consequences and will affect the water quantity and quality offered to downstream users such as the Kruger National Park and to Mozambique. This carries much deeper risks to the system than might be immediately obvious, including the ‘downgrading of the ecosystems services that can be made available in the future. Reducing the river class will also negatively affect the degree to which the system can act as an insurance policy against the impact of adverse conditions (see also Section 5.3). Taking water away from people, i.e. water users, has socio-economic and political implications and is unlikely to be supported by governance structures mandated with water services delivery (Pollard and Du Toit 2011). This option plays the environment off against people – a risky situation where a large number of people rely directly on ecosystem security over the long-run. The most immediate and palatable option to meet the Reserve is to ‘find more water in the system’.

One very feasible option would be to restore degraded riparian ecosystems and land and then manage such a system by applying integrated natural resource management practices such as: maintaining firebreaks, controlling and containing the spread of invasive alien plants to contain them and reducing their extent and density, and ensuring sustainable grazing and browsing regimes. This provides an option for providing a broad-base of social benefits, i.e. all the services linked to the provisioning and maintenance of the ecological Reserve for all people. This implies that the provisioning of the prescribed environmental flows through restoration and management of natural resources does not compromise economic development and the legal tenure of existing water use licenses. It, however, provides an opportunity for government to meet its statutory obligations with respect to the prescribed environmental flows.

One example where such an intervention could work is the Sand River (see Table 5.2). Eight scenarios were investigated, with all but the afforestation scenarios also including the cost of canal repairs (which aim to ensuring a free flow of water to all the users in the catchment):

1. A1.1 Clearing of all forestry plantations in the catchment (deforestation). Irrigated permanent crops (IPC), irrigated annual crops (IAC) and Sabie Sand Game Reserve (SSGR) benefit from increased water flows 2. A1.2 Deforestation. IAC and SSGR benefit from increased water flows 3. A1.3 Deforestation. SSGR alone benefits from increased water flows 4. A2.1 Partial afforestation in optimal areas. Forestry revenue real increase of 7% per annum. Economic value of IPCs is zero 5. A2.2 Partial afforestation in optimal areas. Forestry revenue real increase of 7% per annum. Economic value of IPCs is positive 6. B1 Inyaka dam transfer is completed in accordance with Department of Water Affairs strategic plan capital cost budget 7. B2.1 Operating rule: 35% of water released to the bottom of the catchment, Zone C, with benefits accruing to the SSGR 8. B2.2 Operating rule: 60% of water released to the bottom of the catchment, Zone C, with benefits accruing to the SSGR

The results suggest that, on the whole, positive benefits are obtainable from the improved water supply management scenarios (Scenarios A1.1 to A1.3 and Scenario B1 from the list above). An investment in restoration and catchment management that involves the removal of the remainder of the forestry plantations produces the highest net present values (NPVs) (Scenarios A.1.3). It is also economically viable when irrigated agriculture also benefits from the additional water supply (Scenarios A.1.1 & 1.2). Re-afforestation is not economically viable, even when assuming a 7% real growth in net forestry returns and assuming the value of irrigated permanent crops (IPC) is zero (Scenario A2.1). For the Inyaka dam scenario (Scenario B.1), assuming it is completed within the published DWA investment budget, produces positive NPVs. However, when the specific operating rules (Scenarios B2.1 & B2.2) are accounted for, it generates negative NPVs. Within the context of Scenarios B.2.1 and 2.2 the Maximum Water Demand (MWD) scenario results in a reduction of water flow to agriculture in favour of SSGR, while for the other two scenarios (Allocatable Water Demand (AWD), evapotranspiration (ET)) water flows to agriculture actually increase at the expense of Sabie Sand Game Reserve (SSGR).

Comparing NPV and additional water yields at the bottom of the catchment (Zone C, which benefits SSGR) indicates that the highest potential water yields come through the removal of all forestry (Scenarios A.1.1 to A.1.3). Water yields to Zone C decrease under afforestation (Scenario A.2.1 & A.2.2), with irrigated agriculture the sector most affected by the water losses. Some yield gains at Zone C are obtained through the introduction of the Inyaka dam transfer (B.1). Irrigated agriculture and domestic users are, however, the main beneficiaries. Increases in water yield are higher under the 60% operating rule (scenario B.2.2) as opposed to the 35% operating rule (B.2.1).

Table 5.2 Impact of natural resource management regime changes on net present value and the water flow change in the Sand River Catchment. All acronyms are defined below the table.

Water available: NPV (R million) @ 8% discount end of the catchment Change from baseline rate Zone C (mil m3) (mil m3) Scenario MWD AWD ET MWD AWD ET MWD AWD ET A.1.1 100% cleared 972.54 128.06 336.12 27 38 30 24 0 4 A.1.2 100% cleared 929.71 33.63 263.92 29 38 32 26 0 6 A.1.3 100% cleared 1162.71 111.17 437.62 40 43 42 37 5 15 100% -289.99 -74.19 -381.69 A.2.1 replanted 0 38 17 -3 0 -9 100% -483.76 -120.31 -485.67 A.2.2 replanted 0 38 17 -3 0 -9 100% 204.06 390.16 B.1 completed 7 42 31 4 4 4 B.2.1 35% -92.51 -117.53 -168.73 6 28 15 3 -10 -11 B.2.2 60% -9.13 -100.28 -4.04 10 29 23 7 -9 -3 Source: Crookes (2011b).

Notes: Scenario A.1: removal of all remaining plantations in Sand River Catchment; Scenario A.2: partial afforestation in optimal areas; Scenario B.1. Inyaka dam transfer completed; Scenario B.2. Operating rules established. A positive value for A.2.x indicates that benefits of afforestation exceed opportunity cost. For all other values, a positive NPV suggests that the economic benefits of water supply improvements exceed costs. 7% relates to the real rate of increase in net forestry returns. Operating rule: 35% or 60% water released by irrigated agriculture. NPV values for the B.1 scenario are min and max values (not MWD and AWD scenarios) MWD: maximum water demand (MWD) AWD: allocatable water demand ET: evapotranspiration

Some of the main beneficiaries of an increase in water flow, i.e. meeting the prescribed environmental flows, are the Sabie Sand Game Reserve (SSGR), the Kruger National Park and Mozambique. The question is, should these (and other potential) beneficiaries pay for the delivery of the water? Here we address this by referring only to the SSGR, but the principles could equally be applied to the other entities. The SSGR is a private game Reserve adjacent to the Kruger National Park and caters for upmarket and especially international tourists. The SSGR management has, in the past, contributed towards the clearing cost of a forest plantation containing exotic species in the upper reaches of the catchment (Zone A of the catchment). Due to the mismanagement and ignorance, the water, however, failed to reach the lower reaches of Sand River (Zone C). The SSGR has indicated that they will, again, contribute towards restoration and catchment management, but require some guarantee that the water will arrive in the relevant reaches of the Sand River.

With respect to this case study there are a few challenging questions though, namely:

• As the provisioning of the environment flows prescribed to meet ecological Reserve is a statutory obligation, government has to ensure that flow regimes according to the set river class and Reserve level is maintained – should the SSGR have to pay for it? • If the SSGR pays for the provision of the environmental flows prescribed to meet the ecological Reserve, does that imply that they have an unalienable use right to that water? • Does payment for the environmental flows required to meet ecological Reserve constitute a legal problem? • Does such a payment guarantee the delivery of the service? • Would such a payment create a precedent that would exclude other users from the water in future? Moving towards addressing these questions, we have to recognise that riparian restoration has more benefits to offer than just the volumetric water supply or provisioning services (see Section 5.3 for more details). Although the prescribed environmental flow cannot be used for consumptive extraction, it does not imply that these flows are not without economic value. The environmental flow supports regulatory (e.g. flood damage regulation), cultural (e.g. recreation) and supporting (e.g. habitat) services and some provisioning services (e.g. fish), all of which have some value. It would therefore be fair to say that while the ecological Reserve is a composite of various elements, and the prescribed environmental flow is expressed in volumetric terms, it delivers both volumetric and non-volumetric services. This is an important distinction.

When a financial contribution is made towards the delivery of the environmental flows prescribed to meet the ecological Reserve, it cannot be made on a volumetric basis as the volume can neither be bought nor sold, and is a statutory requirement. Such a financial contribution can, however be made towards the delivery of other water-derived services, such as recreation. If the SSGR is therefore making a contribution towards natural resource management and restoration that will assist in the delivery of the prescribed environmental flows, they cannot pay for the volumetric amount of water per se, but will make a transfer based on the surplus derived from the recreational services they offer as a result of the actual delivery of a volumetric amount of water. Voluntary contributions towards the delivery of a derived service are a transfer – not towards the delivery of the Reserve – which can be considered a form of surplus capture which is made to ensure the delivery of a broader social benefit.

In this case public-private partnerships could be fostered around the restoration and management of catchments and riparian zones whereby both public and private resources can be used to enable:

1. government to fulfil its statutory commitment with respect to the environmental flows required to meet the Reserve without jeopardising economic development and/or entering into a moral debate of water vs. people; and 2. offer an opportunity to establish a mechanism to capture some of the embedded financial benefits (surplus) generated by delivering the environmental flows towards the delivery of a broader social benefit based not on volumes but on ecosystem service delivery.

A key component to the success of such a transaction would be the social contract underpinning the public-private partnership that will provide the necessary guarantees to all concerned that the environmental flows prescribed to meet the ecological Reserve will be provided for the greater good and in line with the spirit of the Act. Such a contract will have to include the party paying for the service (i.e. SSGR), the party responsible for delivering the services (i.e. DWA) and the party responsible for the implementation (i.e. the agency responsible for managing the system and clearing the invasive aliens.) In such an agreement it would be -essential to include performance management indicators and checks and balances, including compensation for losses, in the event of non-performance.

5.5 Message 5: How to deal with unmitigated damage and its costs to society

Sue Milton, James Blignaut and Martin de Wit

5.5.1 Introduction

Environmental legislation provides a mechanism to protect society from the negative effects of development on natural environments and the services that they supply, and if such damages do occur, to manage those. The Environmental Impact Assessment Regulations No. R543 of the National Environmental Management Act (RSA 1998) therefore requires governmental authorisations before environmentally damaging activities may occur. The appropriate authority has to issue a ‘Record of Decision’ (RoD), outlining the criteria by which a destructive activity may proceed. The regulations are very specific about the need to quantify potential damages, losses and risks, and to develop environmental management plans (EMP) that specify how damage and costs to society should be avoided, minimised and mitigated (NEMA 2010: see RSA 1998 amendation). An entity embarking on an environmental destructive activity might also be required to restore the environment following such an activity. In the case of mining, for instance, government may set monitoring and auditing protocols to ensure compliance with restoration targets set in the EMP, before they would issue a mining closure certificate.

Despite these mechanisms and assurances, mining and other developments often reduce the functioning and resilience of ecosystems and thereby reduce the potential of ecosystems to produce goods and services. Such unmitigated damages can often be associated with a net cost for the surrounding community and have a detrimental impact on the future economic prospects for the local economy. Drawing on examples discussed in this report, in combination with a wider literature, we discuss adjustments to target setting, auditing requirements, and ecosystem service amendments that would protect the public from costs incurred by residual damages caused by development.

5.5.2 Target setting (restoration objectives) in Environmental Management Planning

EMP often state restoration targets as percentages of the original species or grazing services to be returned. This was the case with the Namaqua Sands mine EMP drafted in 1990 (EEU 1990) and revised in 2001 (Namaqua Sands 2001) and 2008 (Golder & Associates 2008, Pauw 2011). Namakwa Sands’ rehabilitation goal is to:

...rehabilitate and re-vegetate disturbed areas and establish a self-sustaining Strandveld vegetation cover in order to control dust generation, control wind and water erosion, as well as restore land capability (Golder & Associates 2008).

Short-term targets include reshaping of the mined surface to create a landscape that is resistant to erosive forces such as wind and water. Longer-term targets relate to restoration of biodiversity, natural capital and function of the post-mining landscape. At closure the vegetation should be able to support small livestock (grazing capacity 20 ha / SSU). Three and five year targets for vegetation restoration were expressed as vegetation cover and plant species richness of mined sites in relation to values for references sites (50% and 80 % of the reference site cover at 3 and 5 years respectively) and plant species richness (30% and 60% of the benchmark number of species at 3 and 5 years respectively).

Namaqua Sands has made a major investment in rehabilitation research and implementation, setting a high standard for other mining companies on the West Coast to match. Despite their highly commendable efforts, two practical problems have resulted from the way in which closure targets were defined.

1. The failure to identify functional groups of plant species leads to a de facto net loss of ecosystem values to society even where EMP targets have been reached or even exceeded. For example, the target 60% of plant species returned to mine sites excludes key functional groups such as trees, resprouting shrubs, long-lived shrubs or geophytes. The functional traits of plants matter because annual herbaceous plants, for example, cannot provide the same services as drought tolerant, long-lived, resprouting shrubs in the form of shade, shelter, drought forage, firewood, fruit, or dependable resources for pollinating birds and insects. It is therefore necessary to define restoration targets in the EMP in terms of functions related to self-sustaining ecosystems and the anticipated range of future land-use options. 2. There was no requirement in the EMP that the targets reached or the development trajectories initiated should be maintained over time under a scenario of climatic shocks and grazing management. Pauw (2011) shows that vegetation on Namaqua Sands mine initially increased in cover and diversity on rehabilitated areas, and hence meeting the EMP requirements, but then decreased below benchmark cover levels. Whereas most of the rehabilitated sites achieved the three-year species richness objectives, none achieved the five-year species richness objectives. At ten years after rehabilitation no increase in species richness was detected without additional interventions such as the re-introduction of missing species.

The practical problems mentioned above imply that even after much effort, the productive capacity of the land, both in terms of grazing as well as in terms of biodiversity, has been compromised over a long period of time.

5.5.3 Unmitigated damage

Meeting restoration targets as specified in the EMP does not imply sustainability or that ecosystem function (such as grazing capacity) has been met. This requires a re-think of the way in which the EMP targets are being set, and the way in which the unmitigated damage is being dealt with.

Mugido and Kleynhans (2011) constructed a typical farming model for a 20 year period using the results of Pauw (2011) for both the un-mined and mined yet restored areas of Namaqua Sands. It was found that a farmer has the potential to make more profit farming un-mined land than on rehabilitated land. It is shown that, with a before mining grazing capacity of 10 ha/SSU, the IRR is 13.7% with a corresponding NPV of R3.8million and a BCR of 3.5. In this case the farmer is able to make a profit. After ten years of rehabilitation, the profits from farming on rehabilitated land are significantly lower than the profits of farming on an un-mined land, as the grazing capacity is only restored to, at best, 15 ha/SSU. The resulting IRR is 4.5%, with a corresponding NPV of R377 000 and a BCR of 2.5. The farmer is unlikely to be able to survive as a productive farmer. The balance of evidence presented here suggests a lasting, permanent, unmitigated damage to society as a result of the mining activity despite the best restoration efforts. The real question therefore at stake is how to compensate for this unmitigated loss?

5.5.4 Managing the unmitigated damage to society as a result of transformation of natural landscapes

There are typically two scenarios one has to consider when dealing with the unmitigated damage to society after land transformation has taken place. One is before the RoD has been issued and the other is after the RoD has been issued.

Before the issue of the RoD

Before a RoD has been issued, authorities can consider a range of options to cater for temporal feedbacks to goal setting. This implies the EMP can be written in such a way that the end point of restoration is not given as a fixed parameter, but is equated to the principle that no net loss to society should take place. In the event that there is an unmitigated loss to society, such as in that illustrated by the example above, then further compensatory and offset investments must be considered to ensure that such a loss is addressed and that society is left either in the same or in a better position than before the development has taken place. Examples of such compensatory and offset investments are investments in the carbon market, in biodiversity offsets or in social capital. The Western Cape Provincial Government may require that an ‘Offset Management Plan’, for on-site or off-site offsets be submitted as part of the Basic Assessment Report, or as part of the Environmental Management Plan (DEA & DP 2007).

Arguably the most common offset investment currently is through the carbon offset market. An investment in carbon sequestration is made by which the impact of emitting a ton of CO2 can be negated or diminished by avoiding the release of a ton elsewhere, or absorbing a ton of CO2 from the air that otherwise would have remained in the atmosphere (Ten Kate et al. 2004).

Biodiversity offsets on the other hand are conservation activities designed to deliver biodiversity benefits in compensation for losses, in a measurable way (DEFRA 2012). Ten Kate et al. (2004) define biodiversity offsets as conservation actions intended to compensate for the residual, unavoidable harm to biodiversity caused by development projects, so as to ensure no net loss of biodiversity. Note that before developers contemplate offsets, they should have first have sought to avoid and minimise harm to biodiversity. Biodiversity offsets are thus conservation actions that supplement mitigation and rehabilitation and are generally understood to refer to conservation activities that take place outside the geographic boundaries of a development site or on undeveloped/un-mined parts of a project site. Such offsets should be relevant to local communities. However in some cases, possibly because of a lack of feasible options, biodiversity offsets are made outside the region affected by the development project. A more extreme example would be the funding of a dune or wetland conservation project in a province other than that in which dunes or wetlands were damaged.

Ten Kate et al. (2004) further state that biodiversity offsets have advantages to businesses (improvement of societal trust), governments (spatial development planning), conservation organisations (securing conservation priority areas) and communities (properly rehabilitated project sites, and conservation actions outside the project area that support livelihoods and provide amenities). According to Quétier and Lavorel (2011) there are three possible approaches for offsetting impacts of development projects on biodiversity and ecosystems. All are based on defining ecological equivalents on the basis of assessment of losses and gains for target components of biodiversity (species, their habitat, habitat types, ecosystem functions or services, etc.). DEA&DP (2007) highlights that the actual loss of biodiversity and/or loss or deterioration of ecosystem services that must be compensated needs to be estimated. Two approaches are useful, namely the use of ecological proxies to compensate for residual impacts on biodiversity pattern, process and ecosystem services; and economic valuation of biodiversity to determine appropriate compensation.

In cases where an intervention or extraction of a non-renewable resource will result in a determinable net loss of ecosystem services to society (e.g. long term loss of grazing, cropping, biodiversity or aesthetic values), acceptable investment in offsets and compensation should be discussed and agreed upon with interested and affected parties (IAPS) before the ROD is issued. Potentially acceptable offsets might include investment in reduction of sediment loss, in improved biodiversity conservation or wetland functionality for communities in nearby and somewhat degraded areas that have potential to be improved. Clearly, biodiversity offsets will only achieve results for conservation if they are adequately designed, implemented and enforced (Ten Kate et al. 2004). Moreover, the security of biodiversity offsets must be considered in an offset management plan that confirms that the developer has identified an offset and made provision for management, monitoring, and auditing of the offset, or has made a financial guarantee to the local conservation authority for the securing of an appropriate offset (DEA & DP 2007). Possible achievement or benefits of biodiversity offsets may include:

100

1. Protection of special habitats or areas of intact function or high biodiversity 2. Development of functional ecological corridors between systems through, among others, restoration 3. Increased size, improved shape or increased viability of existing conservation areas Another form of compensation might be through the development of social and human capital in the form of clinics, houses, infrastructure and education facilities. While this is considered common practise, its real contribution to society is seldom considered and hence little attention is given as to how the provision of these services should and/or can make the greatest contribution or their value to the community as infrastructure and employment prospects decay after mine closure. Even when investing in these, its contribution towards ensuring that no net loss to society accrues still has to be investigated.

After ROD negotiations

It is generally difficult to predict the outcome of restoration efforts. Where these fail to meet standards set for project closure, there should be a mechanism for re-negotiating alternative targets with comparable benefits for society. It is generally not in the interests of society to prevent closure of a project because this has social and economic costs in the short term. However, certain standards for closure must be adhered to avoid a situation where premature closure results in long- term risks to health and security. Alternative targets might include offsets as discussed above, or alternative options for clean water supply and food security in the affected area. Such interventions, however, would have to be based on a negotiated outcome following informed discussions between the community, the authorities and the developer. The underlying principle, however, should be that the affected community and/or society should not be left worse off after the development than before. Permanence is an important consideration within sustainable development thinking and hence it is important to ensure that the long-term income and welfare generation possibilities of the affected community are not compromised. It remains, however, important to consider mechanisms to avoid corruption and variation in standards.

As we work and live in a complex and dynamic world, it is not always possible to foresee whether there is likely to be an unmitigated damage even when EMPs have been successfully implemented and targets reached. This implies that after the RoD has been signed, there has to be recourse for the affected community and civil society to discuss ways in which such unmitigated damage can be addressed. In the event that an offset or any other compensatory measure is required, the arguments presented in this paper should be considered.

101

5.6 Message 6: Stakeholder participation: An attempt to close the gap between research and action

Alanna Rebelo & James Blignaut

5.6.1 INTRODUCTION & BACKGROUND

RECOMMENDATIONS: Communication for behaviour change

• The importance of interdisciplinary work: It is imperative that research projects not only investigate the ‘science’ of a system, but that economists and sociologists are involved. Scientists need to learn to communicate with, and work with researchers from other disciplines. All scientific studies should be preceded by a social assessment at the outset to determine needs and values of stakeholders and to determine the capacity of their institutions (Aronson et al. 2010 (2)). • Policy: Scientific researchers should be encouraged, or even required, to publish recommendations for policy and law. • Stakeholder engagement: Scientific researchers should be encouraged and trained to be involved in the communities in which their research takes place, as science cannot be divorced from its social setting. Communities and stakeholders should be involved from the outset and outcomes of the study should be effectively communicated upon conclusion of the research in a fun and engaging manner. • Media dialogue: There is need for a shift in the current approach of the media, where advertising and marketing promotes endless consumption and waste. Furthermore, scientific researchers should be should be encouraged and trained to publish their results in a popular science fashion in the media. This should include all aspects of media: television, radio, magazines and newspapers. • Education: A fundamental change in the education system is required: at all levels, from kindergarten to tertiary education. At high level institutions, students from social, economic, political and natural sciences should be required to learn about each discipline, its basics, its world views, and its assumptions. Universities should no longer be producing single minded researchers working in isolation, unable to communicate or unable to influence decisions at high levels. • Marketing: Communication in itself cannot always inspire a behaviour change. A social marketing strategy should be used to complement other communication strategies.

Ecological, hydrological and economic research in the Kromme River Catchment in South Africa demonstrated that farmers and land-users were using unsustainable management practises that were causing significant damage to the landscape. This was causing deterioration in the supply of ecosystem goods and services of the catchment. The restoration and recovery of the flow of ecosystem goods and services in the catchment would require a change in mindset of farmers and

102

landowners. This would require more integrated communication than simply publishing literature for the scientific community. It includes effectively communicating science to society through direct contact and through the media.

To cement and maintain a behaviour change would require financiall mechanisms, both dissuasive and incentive in nature (Aronson et al. 2010 (1)). This can only happpen by making higher levels of government and municipalities, decision makers and policy makers aware of the issues so that they can make appropriate institutional changes and arrangements to support this. This requires another equally important form of communication: that of convertinng scientific findings into formats easily accessible to engineers, politicians and decision makers.

What is a catchment?

It is an area, often surrounded by mountains, whose rainfall all drains into one body of water, such as the ocean, an estuary, or a ddam. So a catchment consists off mountains, the lowlands/flood plains and all the rivers running through this area. South Africa is divided up very broadly into 19 ‘primary catchments’. Each of these 19 catchments is further divided up into ‘quaternary catchments’, which is the scale at which water management should take place.

Image from: Waterwatch Queensland (http://www.georgesriver.org.au/IgnitionSuite/uploads/images/poster_healthycatchment%20qld%20waterwatch%20sml.jjpg)

103

The story of the Kromme

The Kromme River is a high energy river system in a long narrow catchment in the Eastern Cape of South Africa. It is a very important river as it supplies roughly 40% of Port Elizabeth’s drinking water.

The Kromme River once had pristine palmiet wetlands along the entire length of the valley. These wetlands provided many crucial ecosystem goods and services, including water infiltration (improving water supply, especially during the dry season), water filtration (improving water quality) and decreasing the impacts of floods.

However poor farming and management practises in this catchment have led to the formation of headcuts, or dongas (inset picture), deep gullies resulting in erosion, loss of soiil and a drop in the water table. These headcuts and the invasion of alien trees have destroyed the wetlands and as a result less water is available – especiallyy in the dry seasons, water is of a very poor quality and flood damage is extreme.

Unless this land management is changed in the Kromme immediately, the damage will become irreversible. Restoration of the catchment urgently needs to be undertaken.

To try and effect a behaviour change, a number of channels of dissemination were chosen to ensure effective communication of results and to prompt the appropriate response from stakeholders. These were the following:

1. Direct communication: A workshop was held in the Kromme Catchment for all farmers and landowners to effectively communicate results from the hydrological, ecological and economic studies in the Kromme. This form of dissemination will be expounded upon in this policy brief.

104

2. Indirect communication: Popular magazines favoured by the local farming community were identified. Journalists from these magazines were then approached and asked to feature an article on the results and the stakeholder process. The two magazines targeted for the Kromme case study were Farmers Weekly and Landbouweekblad. 3. Policy change: Municipal decision makers and their advisors were identified. These advisors were sent a presentation of the results at a much more technical level. Meetings were set up to discuss the concepts further. This is with an aim to influence decision makers to take the environment into account by putting protection measures in place and creating financial incentives and dissuasive penalties for landowners.

The Kromme Farmers Workshop

All farmers and landowners in the Kromme were invited to attend a workshop where results of research in the Kromme would be shared, and new studies in the Kromme would be presented. A large amount of effort was made making farm visits to ensure that the workshop was held on a date and time that was convenient for the majority of farmers. Invitations were also delivered by hand. The workshop was held at a neutral, popular, prestigious venue in the catchment called the Plaaswerf. This was done so as to ensure that the farmers and landowners could enjoy the setting, relax and enjoy the feedback. The venue was set up in such a way that it encouraged interaction at all times, including chairs set out in a circle. Many posters and maps were designed such that the information was displayed in a fun, visually attractive manner to compliment the verbal presentations. Conference booklets were also compiled from these posters and other information sheets so that farmers would have this information accessible after the workshop was over.

The workshop was held over an afternoon and ended with a casual dinner and drinks to facilitate community interaction and networking. The programme was designed to be as engaging and interactive as possible, with very short presentations and longer group discussions. A professional workshop facilitator was hired to field questions, manage interaction and direct discussion in an unbiased manner. The programme included a field trip during which two experts spoke for five minutes each about highly relevant topics to the catchment in a casual setting out in the field: fynbos and palmiet wetlands. The workshop ended off with an open session where farmers could air their complaints, concerns and questions.

5.6.2 Analysis: the outcomes

The workshop was extremely successful and most importantly it was enjoyable for all. The farmers and landowners greatly appreciated the effort at communication from researchers and enjoyed the information they received immensely. There were three key outcomes decided at the workshop.

Firstly, the farmers recognized that communication needed to be improved in the area. Secondly, the need to form a committee which would provide a platform for communication among landowners and between landowners and organizations was identified. Most importantly, farmers acknowledged the importance of protecting water supplies, using water more sustainably and the importance of working together towards this goal.

105

The landowners discuss important watter related issues at the Kromme Workshop

5.6.3 Conclusion: the way forward

A key point to make is that “communication in and of itself does not necessariily induce behavioral change, especially when the requirement is tto give up something that is valued” (Aronson et al. 2010 (1)). We recognized beforehand that a behavior change in this farming commmunity will take more than one workshop. As a result a local projeect called “Living Lands” has ‘adopted’ the Kromme River, committing to assist the farmers in the process of becoming more environmentally friendly and managing their land more sustainably. This project has raised funds for the next two years to facilitate this process and will be involved as long as the farmers require their assistance. This highlights the importance of imbedding any research in a more permanent ‘presence’ to be available to providde support and advice if stakeholders are ever to be required to begin this long-term, painful process of change.

106

5.7 Message 7: Policy recommendations towards the acceleration of the restoration of natural capital

Steve Mitchell and James Blignaut

5.7.1 Introduction

The specific restoration intervention required following the degradation of natural capital varies with the type of degradation and the restoration method chosen to combat the degradation. Determining factors include (see also Section 5.2):

• The spatial scale of the degradation, i.e. plot-scale and/or landscape scale; • The degree or intensity of degradation, i.e. slight, severe or the complete destruction of the natural capital; • The type of degradation, i.e. the invasion of alien species, overgrazing, soil loss, open- cast/strip mining, etc.; • The fragility and/or resilience of the degraded ecosystem, i.e. a riparian zone or a desert, an open plain or a mountainous area; • The degree of aridity of the site being restored; • The form of restoration intervention, i.e. more labour or capital intensive and whether the restoration is intensive (active) and/or less intensive, even to the point of being passive; and • The biotic/abiotic nature of the degradation, i.e. whether the degradation only affects the biotic aspects of an ecosystem (plant and animal life), and/or whether it also affects the abiotic aspects (e.g. soil structure, soil loss, etc.).

Both the risk profile as well as the likely financial and economic return on the investment of any restoration project is heavily influenced by the variability. The standard and most commonly used portfolio map in financial theory is the risk-reward bubble plot (Figure 5.2), with the size of the bubble indicating resources committed to it. This provides the means of comparing projects by considering a range of factors (reward or payoff, probability of success, and cost). Using this two-by- two classification system one can differentiate among four possible outcomes when comparing the probably of success (from low to high) as measure of risk and the reward (from low to high) as a measure of return. These four possible outcomes are, conventionally, called as follows: Oyster (high risk projects with uncertain merits), Pearl (projects with high likelihood of success), Bread and Butter (essential projects that enterprises cannot do without), and White elephant (projects which are preferable to avoid) (see Chapter 4).

When considering risk, however, it has been established that no one or single measure of risk (resources committed, standard deviation or coefficient of variation) is sufficient for selecting and classifying projects. All three risk analysis determinants are required in a combination to provide an improved means of selection (Figure 5.2). The data from the resulting analyses can then be used to inform a portfolio mapping exercise, in order to classify and select restoration projects (Table 5.3). A

107

summary of the classification of projects suggests that the projects with the highest potential payoffs (and therefore are pearl projects) are the water projects, in other words those projects where downstream water consumers benefit from the restoration project. Agulhas, Beaufort West, Kromme and Sand are all examples of this.

Table 5.3 Summary of projects classified by type

Oyster Pearl Bread and Butter White elephant High risk projects Projects with high Essential projects Projects which are Description with uncertain merits likelihood of that enterprises preferable to avoid success cannot do without Water projects Drakensberg; Ag, BW, Kromme Kromme (no (with agriculture), agriculture) S) Crop projects Sabie Sand Agulhas, Kromme (with agriculture) Grazing Lephalale Oudtshoorn Beaufort West, Namaqua Sands projects (passive only) Drakensberg, Kromme (with agriculture)

The results indicate that water projects, while individually performing the best, are not sufficient on their own to mitigate the risks of the project. Table 5.3 shows that those projects that include agriculture (in the mix) are subject to lower risk. Firstly, Kromme without agriculture is classified as oyster (in other words, more risky) compared with Kromme (with agriculture), which is classified as a pearl. Secondly, in the Sand study, in the case where Sabie Sand Game Reserve only benefits from the water is a higher risk project than restoration where irrigated agriculture also benefits. Another restoration study which is too reliant on water for benefits is the Drakensberg study, which is also classified as an oyster. Communal agricultural benefits and carbon values are not sufficient to increase resilience in the system. Lephalale on the other hand, is too reliant on grazing, and the introduction of a biomass electricity plant could potentially mitigate that risk and even push the project into an oyster or bread and butter project.

The bread and butter projects are mostly -cropping or grazing projects, but these are only profitable if combined with either water or biomass projects. The bread and butter projects are examples of Smithian innovation, where the division of labour results in qualitative improvements in outcomes. These project benefits are essential to ensure the success of restoration activities. A diverse project portfolio requires both Koestlerian innovation and Smithian forms of innovation.

The analysis of projects using portfolio mapping suggest that this approach, coupled with risk analysis and system dynamics modelling, is able to provide a means of selecting and prioritising restoration projects. Caution should however be exercised in interpreting the results. A positive NPV should not be interpreted as a license to exploit the natural environment. Another area where caution should be exercised in applying the results of this study is in the case of critical natural capital. In such cases restoration should proceed regardless of whether or not NPVs are positive (see e.g. Farley and Gaddis, 2007). This study does however indicate that economic criteria are not the

108

only norm for determining restoration sites. Other criteria that consider ecological, technological and process risks are also important.

1.2

0.8

0.6 Grazing Water -100000 -500000.4 0 50000 100000 150000 Crop

0.2

-0.2

1.2

0.8

0.6 Grazing Water -100000 -50000 0.4 0 50000 100000 150000 Crop

0.2

-0.2

1.2

0.8

0.6 Grazing Water -100000 -500000.4 0 50000 100000 150000 Crop

0.2

-0.2

Figure 5.2 Integrated portfolio map for different ecosystem services (from top to bottom, bubble sizes indicating: Resources committed, standard deviation, coefficient of variation) (S = Sand river, D = Drakensberg, Lp = Lephalale, N = Namaqua Sands, BW = Beaufort West, Ag = Agulhas, Ka = Kromme Agric, Kna = Kromme no-agric, Ou = Oudtshoorn)

The spontaneous development of markets for ecosystem goods and services is likely to occur only where the risks are low and the returns high. In the other three sectors (Figure 5.2) the development of markets for ecosystem goods and services is constrained. There may be select opportunities for venture capital investment where the risks are high and the returns are high. The likelihood of only finding public investments in cases with low risk and low returns is high. In cases where there is high risk and low returns, the recipients must be cautious of any activity that might lead to degradation in

109

the first place, and put preventative as well as adaptive measures in place. Under arid conditions, for instance, the process of restoration is likely to take a long time, with the return on investment taking decades to materialise, with associated high risk.

Based on the above, we commence with a few general policy recommendations followed by a number of specific recommendations.

5.7.2 General policy recommendations and some observations

The following generic policy recommendations are made based on the research conducted in this project:

Current restoration initiatives

Restoration is currently widely undertaken in South Africa, with much of it being conducted through the Natural Resource Management projects of the Expanded Public Works Programme. Large areas have been cleared of invasive alien plants and the follow-up is generally good, but the areas are not always seeded with indigenous plants to replace the cleared aliens. Relying on the existing seed bank may mean that the regeneration of indigenous flora may be slower than it need be.

This research has, in each case, taken the restoration to an appropriate end point and methods for assessing the success of each intervention have been proposed and tested as far as possible within the limited duration of the project.

Recommendation:

The methods used to take rehabilitation to an appropriate end point and to measure the success of the intervention could be applied more widely through other projects, potentially improving their success rate in terms of considering risk and reward as illustrated above.

Restoration should be anchored in the local economy

In each of the case studies conducted during this research, the researchers interacted closely with the land owners and local communities. The results reflect the input of these stakeholders who have understood what the research was aiming to achieve. This is an important step in ensuring that the results have a good chance of being used. In the Kromme River case study the research team has spent time explaining the outcomes to the farming community after the main project was over. This has ensured that there is a high level of understanding of the work.

110

Recommendation: 1. Owners or users of land on which restoration interventions are being undertaken are conversant with and supportive of the restoration in order to motivate their on-going support and maintenance of the work. This does, however, require high-level buy-in by the authorities. 2. The beneficiaries of the restoration will contribute either in part or in full depending affordability towards the costs of the restoration work.

Temporal and spatial dynamics matter (e.g. initial lag, time to full effect)

In each case there is a time lag between the restoration intervention and the realisation of the benefits of the interventions. This may range from less than a decade in the higher rainfall areas to several decades in the arid areas.

In addition, in some cases the benefits accrue to the area where the intervention has been made, but benefits can also accrue to off-site users. These may be downstream, but in instances such as the Kromme River, the Nelson Mandela Metropolitan Municipality, which is located in a different catchment, could be the beneficiary.

The lag period and the spatial distribution of the benefits have implications for the funding and management of these interventions. In general, restoration should be funded by the beneficiaries of the interventions.

Recommendation: Restoration interventions should be planned and executed knowing who will be the beneficiaries of the activities, and these beneficiaries should be willing to pay for the benefits of the restoration process. This is particularly important where occupiers of the land may be expected to play a part in maintaining the interventions while receiving only part of the benefit from the restoration or during the lag period before the benefits of the restoration are realised.

Return on investment

The studies have provided methodologies for assessing the level of risk and the return on funds invested in restoration. There is a clear relationship involving the trade-off between risk and return and where investors are likely to go first. Private involvement may be sought in situations where a high return and low risk is expected. But in situations where the return on investment will be lower or the risk higher, and where restoration is needed for other than economic reasons, then government, as custodian of the public good, is most likely to be the main investor.

When considering the development of markets and payments for ecosystem goods and services this dynamic is to be considered.

111

Recommendation: The costs of restoration should be apportioned to the beneficiaries, taking into account the risk profile and return on investment as illustrated in Figure 5.1.

Technical – environmental science-based aspects

Only some ecosystem services are traded in the market and so the economic value of some services is often not recognised. There is a need to consider the full suite of ecosystem services which will be affected in order to avoid unexpected consequences. Consider also Section 5.3 in respect of the aspects that have to be considered when assessing restoration projects.

Recommendations:

• Create capacity to understand the national need for restoration and also what can / needs to be done at both the national and the local levels.

• Understand the concept of natural or ecological capital (or infrastructure) and how this contributes to the socio-economy

• Natural capital needs to be added/injected into the budget along with man-made or manufactured capital

• Planning any intervention should take account of the position in the landscape in relation to the kinds of services being addressed

Governance

National policies and legislation (e.g. NWA, NEMA, CARA) provide the framework within which restoration projects are supported. Planning and decisions for individual restoration projects should be based on a participatory approach, involving users, planners and policy-makers at all levels, as appropriate. This will contribute to the sustainability of the projects. The tenure system should be taken into account, with principles of equity and redress providing input into the process. Markets for the outcomes of the restoration should be considered in the initial planning phase.

Recommendations: • Land owners / occupants are committed to work with the restoration and maintain it as necessary; • Government stimulation may be required in the absence of existing markets; • Policy and legal goals and imperatives, where these apply, provide the overarching framework within which restoration will be carried out; • Capacity / capability for the implementation of legislation should be developed where this is weak.

112

Institutional

Where there is private or government ownership of land or resources the institutional arrangements within which restoration projects are implemented are clear, but where tenure is communal or open access then the institutional arrangements will need to be clarified and supported. The relationship should be formed between the occupants of the land where the restoration will occur and the beneficiaries where these are off-site to ensure that the arrangements for the sustainability of the project (PES, Equity, Redress, etc.) are enshrined.

Recommendations:

• The costs and benefits (both public and private) of the restoration intervention are defined and understood by all parties involved

• Where the beneficiaries of a restoration intervention are off-site it is recommended that the relationship between the beneficiaries and the occupants of the land is formalised in a way that will protect the interests of all parties

• Occupants and beneficiaries should all work towards the same goal.

• Making provision and/or developing the means and the mechanisms whereby local communities and/or conservation agencies can raise the funds for restoration work through a Green Fund or the like.

Following from the research conducted and the policy recommendations above, a few general observations can be made, namely:

Costs and benefits

All direct and indirect costs and benefits associated with different land-use options must be identified and compared. If the cost or benefit is not included in an explicit market, or if a monetary value for the attribute does not exist, it must be included in a qualitative manner.

Distribution of benefits

It is necessary to understand how the demand and supply of ecosystem services varies within an area as well as accounting for downstream (off-site) benefits.

Scenarios

Different scenarios of how ecosystem services can be expected to change over time must be envisaged, and livelihood options selected accordingly.

113

Objectives identified

Objectives of any restoration must be clearly identified.

Location

The selection of the restoration site is crucial, as this affects the supply of environmental goods and services to the market. Features to consider include the extent of degradation, distance from the market, the existence of an appropriate distribution channel, and the nature of the land tenure. Other factors include the climatic conditions, and also the vegetation type, as this will affect the range of environmental goods and services supplied. There is a need to start in prioritised areas where there is a greater chance of success.

Functionality

Restoration has the aim of restoring the functionality of an ecosystem. This in turn affects the flow of goods and services onto the market. However, this can only occur if markets themselves are well functioning, where possible free from restrictions and characterised by many buyers and sellers. Where markets are not well functioning, some monitoring is required to ensure that dominant players do not manipulate the system to the detriment of other participants.

Process

Many environmental commodities are traded as free goods, or at a price that does not reflect their true value. One example in South Africa is water. A key market challenge for restoration is to determine an appropriate price for environmental goods and services. A number of techniques such as those discussed here in this report are available in order to achieve this.

Protect what we have

Protecting our current natural resources will ensure their services are delivered in perpetuity. The Kromme case study, which illustrates this, may be applied to other catchments in South Africa.

The cost of doing nothing is high

If restoration is not effectively done, catchments will become increasingly damaged and the derived benefits will decline proportionately.

It is more cost effective to maintain natural infrastructure in good condition than it is to restore it. Every effort should be made to prevent degradation in the first place.

Legislation

Existing legislation should be modified to regulate the process of restoration to ensure that the desired end-points are reached.

114

• Acknowledging that restoration may take a long time, there needs to be sufficient flexibility to ensure that end points are met under the highly variable conditions that occur in South Africa. • The introduction of incentives for the polluter to comply beyond minimum requirements of permits/licenses could be introduced.

5.7.3 Specific policy recommendation

This research consists of eight case studies on the restoration of natural capital, focusing on water and agriculture (Table 5.4). Each of these studies centred on a particular issue relevant to the area. The areas studied ranged from a high rainfall area in the Mpumalanga escarpment to the hyper-arid west coast.

115

Table 5.4 Main findings and recommendations of the eight restoration sites

Study site Biome Issues Recommended intervention Main findings Specific recommendations Namaqualand Strandveld Strip mining disrupts natural ecosystems, • Topsoil is removed before mining, the • The EMP does not go far • The requirements of the (Brand-se- removing all vegetation and changing the tailings are spread in the mined areas enough to ensure recovery EMP should allow for Baai) topography, soil structure and chemistry. and the topsoil replaced. of the veld intermediate end points The EMP requires that grazing be • Wind-breaks are erected • The first plants to colonize in the restoration restored to 20 ha per small stock unit. perpendicular to the wind direction. the restored land are the Setting of the restoration goals according • Restoration techniques used:- varying less palatable ones to the EMP seems to be inadequate the topsoil replacement, seeding and leading to an unmitigated loss after plant translocation. restoration • Reconsider the EMP goal setting process Nama Karoo Nama Invasion by alien plants, with the • Prosopis cleared by a Working for • Deep-rooted Prosopis • After clearing the (Beaufort Karoo mesquite Prosopis sp. being the worst. Water team. constrains ground water Prosopis, the land should West) The deep root system is thought to have • The internalisation of the marginal recharge be seeded with reduced the availability of groundwater value of water following restoration • Clearing Prosopis improves indigenous plants and both through intercepting water in the and natural resource management as grazing capacity allowed time to establish upper levels and the deep roots tapping an water supply augmentation option • Return on investment before being grazed the groundwater. This water forms part small pending the of the Beaufort West water supply. marginal value of water Little Karroo Succulent High densities of ostriches have caused • Hand-dug hollows (0.25 m deep by 1 m • Seeds germinated in hand- • Results are used to (Oudtshoorn) Karoo severe damage through trampling, diameter) to retain rainfall runoff and dug hollows successfully inform farm-specific destroying vegetation and exposing seeds. These grew significantly higher • While the study showed biodiversity best topsoil to erosion. The soil surface densities of Tripteris sinuatum and recovery of less palatable management guidelines becomes compacted, rain does not Ruschia spinosa than the other species after 18 months • Government to infiltrate the soil, dongas form on the treatments. with active restoration stimulate the move to paths and seeds are unable to germinate. • Tractor-drawn ripper to break the methods, recovery of the small-camp system Unsustainable ostrich farming practises crust and allow rainwater infiltration more palatable species through PES. This would have an impact on the marketing and and roots to penetrate. may take more than 2 need interventions at branding of the products forthcoming • Seeding* – see table below decades. national policy and local having an economic consequence • Mulching • Small camp system gives levels. • Re-design farming practises to become higher return and more sustainable and allow eco- preserves the veld branding and international marketing • Small camp system can support active restoration

116

of products Lephalale Savanna Changes in land use have contributed to • Clearing of trial plots to measure the • Bush thickening reduces • The large trees should bush thickening, reducing the quality of change in the quality of grazing and the carrying capacity of be maintained during grazing available. browse. grazers and ultimately of bush thinning • The optimum tree density is 4,000 browsers as well • Leguminous trees fix ETTE†, bush-thickened areas can reach • Thinning becomes cost nitrogen and so provide 12,000 ETTE. effective when the wood islands of richer grazing • The potential for the development of a is sold on the informal under their canopies renewable energy (biomass-based) market sector should be actively pursued • Develop renewable energy market Agulhas Plain Fynbos Invasive vegetation has transformed • Alien invasive plants removed and • Estimated 82 million m3 of • Impact of fire on aliens much of the fynbos, leading to a seeds of indigenous fynbos sowed, water will be made in fire-adapted fynbos reduction of the natural capital such as • Managed rotational burns according to available through clearing needs research fynbos flowers for export. optimal fire frequency, i.e. once every of alien vegetation • Develop markets for PES 15 years • The value of the water will to incentivise • The internalisation of the marginal cover the costs of clearing conservation and value of water following restoration alien invasives restoration of natural and natural resource management as • Flower harvesting on its capital an water supply augmentation option own will not cover costs of alien eradication Kromme River Riparian Overgrazing and other agricultural • Removal of alien plants • Turbidity decreased (not • The programme should zone practices have led to widespread erosion • Rehabilitation of wetlands through significant) with wetland be clearly communicated and loss of wetlands. Widespread installation of gabions, soil restoration, although and everyone should invasion of alien plants. stabilisation and revegetation there might be a lag effect work towards the same There is a disconnect between the • Development of the necessary • WfW not viable when goal landusers (farmers) and the main users of institutional mechanisms to connect assessed only on flow • Development of a PES water (the municipality) the landusers to the users of water increase system Drakensberg Montane Overgrazing and other agricultural • Removal of alien plants • Infiltration higher on • The programme should (Okhombe grassland practices have led to widespread erosion, • Rehabilitation of grasslands through restored plots be upscaled and village) poor veld condition, and the reduction of soil stabilisation through the • Passive restoration developed across many water infiltration. installation of gabions, packing stones, unlikely to succeed Drakensberg catchment making swales, and revegetation • From the private areas and include • Development of the necessary landowner’s perspective considerations for institutional mechanisms to connect rehabilitation does not auxiliary programmes the landusers to the users of water pay, but the reduced focusing on the

117

and soil stabilisation services siltation alone ensures development of Green that rehabilitation is viable Villages††, i.e. food, from a societal water and energy perspective. security measures • Species richness on • The development of a restored plots showed no market for soil trend, being better at one stabilisation services site poorer at another and unchanged at two others. Sand River Savanna/ Non-delivery of the environmental • Refurbishment of existing water • There are various water • Restoration and the Grassland reserve mainly due to plantation forestry infrastructure augmentation options to removal of the and the degradation of the water supply • Establishment of a natural resource supply the instream remainder of the infrastructure management area in the upper- waterflow and ecological afforested area to The failure of the water resource catchment, which could include reserve requirements; commence as soon as management institutional system and activities such as the removal of • The cheapest and possible hence the inability to guarantee the invasive alien plants and integrated economically most viable • Funding of the delivery of the environmental reserve catchment management option to supply the restoration work could • Development of required institutional instream flow requirement entail a public-private mechanisms to link the lower is through the reduction of partnership catchment water users to the water the afforested area. users in the upper parts of the catchment

118

*Species of seed sown – Morestȇr Farm, Little Karoo.

Seed mix Quantities Total no. of Growth form Species Kg seeded seed/m2 seed sown Grass Fingerhuthia africana 0.2 54,560 17.96 Succulent Ruschia spinosa 0.1 40,000 13.17 Succulent Drosanthemum hispidum 0.2 280,000 92.18 Shrub Tripteris sinuata 0.4 24,880 8.19 Shrub Tetragonia fruticosa 0.3 8,520 2.80 Shrub Lessertia annularis 1 144,800 47.67 Total (mean/m2) 2.2 552,760 36

†ETTE deﬁnion: “tree with the same leaf volume as a single stemmed 1.5 meter tree” (Cloete, J, 2011, pers. comm.)

†† Green village: a green village in this context refers to a sustainable village from a food, water and energy perspective that may or may not function independently, i.e. off-grid, with a low ecological/environmental footprint.

119

5.7.4 Conclusion

Ecosystem services provide essential, though often invisible (and thus unrecognised), contributions to the socio-economic system and the functionality of ecosystems needs to be maintained in the face of increasing demands. The variety of conditions and interventions studied under this research indicate that restoration is economically feasible in most cases, but with varying risk profiles and the return on investment will also vary depending on mainly on restoration costs and the marginal value of water. This will, to a large extent, determine the source of the funding necessary to undertake the interventions. It is generally less costly to maintain ecosystems in good condition than it is to restore them, and the contribution of ecosystem goods and services is cardinal to our wellbeing. It is important that EGS are accorded the value that they deliver. In this way ecosystems will continue to deliver the stream of goods and services that we, often unconsciously, rely on.

120

CHAPTER 6 CONCLUSIONS, RECOMMENDATIONS AND FUTURE RESEARCH

Steve Mitchell, Martin de Wit, James Blignaut, Sue Milton, Karen Esler, David le Maitre, and Douglas Crookes

6.1 Introduction

The focus of this study was on conducting a meta-analysis, of the hydrological, ecological and socio- economic impacts of restoration at eight different sites primarily from an agricultural perspective. The meta-analysis comprised two components, namely:

1. a literature review (considering over 20,000 peer-reviewed published papers) focussed on the question whether published work on restoration incorporates the socio-economic and policy implications, and 2. the construction of a systems dynamics model synthesising the socio-economic and biophysical information from eight existing and ongoing restoration projects into a shared understanding. The model was based on empirical research of the ecological, hydrological and economic impacts of restoration, as structured in several Masters theses’. This was followed by an integrated economic evaluation over all sites with the assistance of the RESTORE-RT model, a systems dynamic model developed in a PhD thesis for this research project.

To recap, the aims of the project were provide the information needed to test the project’s hypothesis in Box 6.1.

Box 6.1: The hypothesis addressed by the research:

The aims have been fully addressed and were:

121

2. To conduct hydrological assessments to determine the impacts of restoration at eight ecologically and socio-economically different restoration sites in comparison to degraded or unrestored areas in close proximity to the eight selected sites 3. To conduct economic and socio-economic assessments to determine the impact of restoration at eight socio-ecologically and socio-economically different restoration sites in comparison to degraded or unrestored areas in close proximity to the eight selected sites 4. To investigate the impact of restoration on sustainable rural employment and payment for environmental services 5. To develop a user-friendly model that could be used in future to model the likely impact of restoration of the ecology , hydrology and economy – notably agriculture. 6. To prepare a meta-analysis as a synthesis of all the studies above, based on the outcomes of the research.

To achieve these aims a very extensive literature review was conducted and is documented in Chapter 2. Chapter 3 focuses on the research method with specific reference to system dynamics modelling. Chapter 4 contains the case study and modelling results with Chapter 5 reflecting on some of the key messages and emerging issues based on the research. This chapter synthesises the findings and concludes the report. The information in the report is supported by a number of Annexures. Abstracts of the student projects are given in Annexures A to I. The full theses of the students are available on CD and from the various university websites.

6.2 Main conclusions

It is important to note that the project focused on the economics of restoration, specifically on the returns versus risk trade-offs over time. The research generated a wealth of information on ecological and hydrological responses before and after restoration activities, but in the synthesis this information was used primarily to provide an economic evaluation of restoration, and more specifically to evaluate both the economic risks and economic returns for each restoration site. The study did not investigate the need for restoration, nor did it consider any ecological threshold analysis to determine when restoration becomes critical. The results should also be interpreted within the context of current/existing data as pertaining to the existing restoration projects. The plausible impact of any technological advances, and/or improved restoration methods, or even changes to the restoration cost structures has not been considered.

122

Table 6.1 Summary of projects classified by type of risk and the likelihood of returns at each of the eight study sites from a market for ecosystem goods and services’ perspective*

Oyster Pearl Bread and Butter White elephant High risk projects Projects with high Essential projects Projects which are Description with uncertain likelihood of that enterprises preferable to avoid merits success cannot do without Water projects Drakensberg; Agulhas, Beaufort Kromme (no West, Kromme agriculture) (with agriculture), Sabie Sand Crop projects Sabie Sand Agulhas, Kromme (with agriculture) Grazing projects Lephalale Oudtshoorn Beaufort West, Namaqualand (passive only) Drakensberg, Kromme (with agriculture) Note: * a project being non-desirable from a market for ecosystem goods and services’ perspective does not imply that restoration per se is undesirable. It still might be desirable, even essential, but for non- monetary reasons.

The bread and butter projects are almost entirely crop or grazing projects, but these are only profitable if combined with either water or biomass projects. These water and biomass benefits are essential to ensure the success of restoration activities.

In general, economic evaluation of these restoration case studies shows that the costs of restoration exceed the private benefits realised through restoration, but that when social benefits (for example...) are included, they do cover the costs under certain conditions. Two case studies that support this thesis are the Kromme River and the Agulhas Plain. The restoration cost associated with the water made available through restoration activities is comparable to the price of water in the urban areas served (e.g. Nelson Mandela Metropolitan Municipality and Bredasdorp). Private benefits are mainly related to improved grazing for livestock and/or game, and biomass energy production under certain conditions, including the kinds of production activities that are appropriate for those environments.

The high cost of restoration together with the (sometimes significant) time lag before benefits are realised makes restoration activities unprofitable for private entities in most cases. Expert opinion is that time lag between restoration and the realisation of benefits in arid areas such as Oudtshoorn and Namaqua Sands amounts to several decades. Based on an earlier long-term study in the Lephalale area, the time lag is over a decade. However, the value of the water that becomes available following restoration is more immediate and, in most of the cases studied, is what makes restoration economically viable.

123

The results are particularly sensitive to restoration costs and the marginal value of water and less sensitive to the value of the benefits derived from other ecosystem goods and services. The importance of variation in climatic factors can only be affirmed by research over longer time frames.

The literature review (Ch. 2) indicated that the link between restoration of natural capital and the benefit it generates for society are not well reflected in the literature. The majority of papers were from developed countries and focus on ecological and biodiversity aspects. The link between restoration and payment for ecosystem services was addressed in a very small proportion of the published studies. Restoration studies from low income countries had a greater focus on livelihoods and often made the connection between livelihoods and ecosystem services such as the provisioning of materials/goods or regulating services.

Ecosystem services can be restored, but this may take a long time. Expert opinion is that in the more arid areas, the lag between restoration and the realisation of the restored services may take several decades. The study itself did not provide proof for such an argument, but presented a snapshot of the present situation at each site which indicates that the recovery is slow.

6.3 Recommendations

Based on this research the following recommendations are made:

Technical / operational recommendations

Use of conventional economic evaluation tools: This study showed that conventional economic evaluation tools may be used for the valuation of the restoration of natural capital.

124

(manufactured or man-made) investments which are a means to an end and aimed at delivering a vital public good or service rather than a purely private one.

We recommended that restoration of natural capital is recognised as a water supply option and needs to be seen as an investment in ecological infrastructure that yields, among others, water services.

Restoration as an insurance against risk: There is a pressing need to supplement financial insurance with nature-based insurance policies which provide a hedge against adverse impacts on natural capital. The similarity between financial insurance and ecosystem-based approaches has been recognised (Van Oosterzee et al. 2012). Ecosystem services themselves, however, have no recognised hedge mechanism. We proposed adapting van Oosterzee et al. (2012) approach to hedging against risk in a REDD (Reduction in Emissions through Degradation and Deforestation) programme, to provide a mechanism hedged against a failure or malfunctioning of natural systems, and thus the reduced or non-delivery of ecosystem goods and services, as a result of degradation and a loss in quality. We suggest that restoring the functionality of degraded ecosystems will offer a bio-physical, system-wide, insurance against the effects of adverse events. Ecological restoration activities to restore resilience are therefore likely to safeguard against the loss of essential services, as well as against the consequent loss of manufactured, cultivated and human capital.

125

We recommend that, where water is needed to meet the requirement of the ecological reserve, the option of restoring the natural capital be investigated. In doing this, consideration should be given to the questions raised in Section 5.4.

Restoration should be anchored in the local economy: In, among others, the Kromme River case study the research team spent time explaining the outcomes to the farming community after the main project was over. This has ensured that there is a high level of understanding of the purpose and benefits of restoration work.

126

We recommend that the beneficiaries of restoration contribute, either in part or in full, towards the costs of the restoration work.

We recommend that the costs of restoration should be apportioned to the beneficiaries, taking into account the socio-economic and risk profile and return on investment.

Technical – environmental science-based aspects: Only some ecosystem services are traded in the market and so the economic value of some services is often not recognised.

Recommendations:

• Create capacity to understand the national need for restoration and also what can or needs to be done at both the national and the local levels;

• Develop a greater understanding of the concept of natural or ecological capital (or infrastructure) and how this contributes to the socio-economy;

• Natural capital needs to be valued and these values added/injected into the budget along with man-made or manufactured capital;

• Planning of restoration interventions should take account of the position in the landscape in relation to the kinds of services and benefits being addressed.

127

Recommendations:

• Ways need to be found to ensure that land owners / occupants are committed to work with the restoration and maintain it as necessary;

• Government stimulation may be required in the absence of existing markets;

• Policy and legal goals and imperatives, where these apply, provide the overarching framework within which restoration will be carried out;

• Capacity / capability for the implementation of legislation should be developed where this is weak. Institutional: Where there is private or government ownership of land or resources the institutional arrangements within which restoration projects are implemented are clear but, where tenure is communal or open access, then the institutional arrangements will need to be clarified and supported.

Recommendations:

• The costs and benefits (both public and private) of the restoration intervention are defined and understood by all parties involved;

• Where the at least some of the beneficiaries of a restoration intervention are off-site it is recommended that the relationship between the beneficiaries and the occupants of the land is formalised in a way that will protect the interests of all parties;

• Residents of the site and beneficiaries should all work towards the same goal;

• Make provision for, and/or develop the means and the mechanisms whereby local communities and/or conservation agencies can raise the funds for restoration work through a Green Fund or the like.

Future research

The following is a non-prioritised list of questions for future research that stems both directly and indirectly from the research conducted.

128