A Re-Analysis of the South African Catchment Afforestation Experimental Data

A RE-ANALYSIS OF THE SOUTH AFRICAN CATCHMENT AFFORESTATION EXPERIMENTAL DATA DF Scott • FW Prinsloo • G Moses M Mehlomakulu • ADA Simmers WRC Report No. 810/1/00 Disclaimer This report emanates from a project financed by the Water Research Commission (WRC) and is approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the WRC or the members of the project steering committee, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. Vrywaring Hierdie verslag spruit voort uit 'n navorsingsprojek wat deur die Waternavorsingskommissie (WNK) gefinansier is en goedgekeur is vir publikasie. Goedkeuring beteken nie noodwendig dat die inhoud die siening en beleid van die WNK of die lede van die projek-loodskomitee weerspieel nie, of dat melding van handelsname of -ware deur die WNK vir gebruik goedgekeur of aanbeveel word nie. A RE-ANALYSIS OF THE SOUTH AFRICAN CATCHMENT AFFORESTATION EXPERIMENTAL DATA Report to the Water Research Commission by DF Scott, FWPrinsloo, G Moses, M Mehlomakulu and ADA Simmers CSIR Division of \Yater, Environment and Forestry Technology Stellenbosch CSIR ENVIRONMENTEK CSIR Report No. ENV-S-C 99088 March 2000 WRC Report No 810/1/00 ISBN No 1 86845 672 2 EXECUTIVE SUMMARY INTRODUCTION collected, the length of record and the range of sites involved make the experiments Forestry is an important agricultural activity in unique and invaluable. Although most of the South Africa accounting for 6.3% by value of experimental data have already been the country's gross agricultural output in analysed to some extent or another, there 1996/97 (FOA, 1998). This primary industry was a need to consolidate the experimental in turn supports a large forest products data, to re-work the data in a uniform and industry that made up 7.4% of South Africa's consistent way, and to generalise the results, gross manufacturing output in 1996/97 (FOA, particularly with respect to new information 1998). At the same time forestry is estimated needs. We can now review this large body of to have a consumptive water use equivalent data to assess the composite picture which to 7.5% of the country's available water has emerged to date with respect to the resources (Scott et at., 1998), and forestry influence of forestry on streamflow. continues to receive much attention because of the water consumption of timber This project for the Water Research plantations. The expansion of the forest Commission sought to:- industry has been regulated since 1972 on • Update and prepare the complete body the basis of its estimated water resource of experimental afforestation data. effects. The provisions of the new National • Produce a definitive review of the South Water Act will make forestry liable to pay for African catchment afforestation water use as a streamflow reduction activity, experiments based on a complete and and this has placed increased attention on the consistent analysis of the available data. quantification of the water use of timber • Consolidate our understanding of the plantations, and the accuracy with which such nature of and our ability to predict use can be estimated. For this reason it is streamflow reductions caused by important that full use is made of the forestry. information contained in the long-running • Provide a definitive baseline against catchment afforestation experiments. which modelling efforts can be tested. The South African afforested catchment METHODS experiments were initiated over sixty years ago, with a long term vision of multiple The study looked at a series of paired replication of sites and species over many catchment comparisons. This method decades. The historical nature of the data involves the long-term monitoring of A Re-Analysis of the South African Catchment Afforestation Experimental Data Executive Summary, page i streamflow from pairs of catchments before are standardised to a 10% level of planting or and after a major vegetation change in one of clearing and plotted against time in two them. The treatment effect is primarily figures. The seasonal effects are illustrated measured against a baseline provided by the by plotting the mean flow reductions or relationship between the two catchments increases for each month of the year, before treatment. The method is applicable to generated over many years while the both afforestation and clearfelling treatments. plantations were mature. The results are also The statistical test for treatment effect is by tabulated. means of the dummy variable technique of regression analysis. The initiation of flow reductions (onset of significant reductions after planting) varies Weekly streamflow volumes were used as the widely depending on the stature of the computational unit, which provides some competing native vegetation and the rate at smoothing when comparing different which catchments are dominated by the catchments, but at the same time provides for plantation crop. The pine plantations in tall robust statistics because of the large sample fynbos in the Western Cape and in high numbers. The effects on both total flows and altitude grasslands in the Drakensberg low flows were analysed separately: low flows usually took several years to have a clear being defined as those weeks when flow was impact on streamflow {up to 6 years). below the 75th percentile exceedance level in However some pine crops, e.g. the control catchment. Lambrechtsbos-A in Jonkershoek and Mokobulaan B in Mpumalanga had an early Seventeen experiments were analysed effect on streamflows (within 3 years). altogether, from data generated in 13 treated Eucalypts have an earlier impact on catchments, and comprising 12 planting streamflows, within 2 to 3 years. Under drier experiments and five clearfelling experiments; conditions this was still true, though here 12 of which experiments were with pines and (Ntabamhlope) the timber crop also had the five with eucalypts. The research sites are at benefit of full site preparation prior to planting. five locations across the forestry (i.e. high rainfall) zone of South Africa, namely, Once reductions are significant they generally Jonkershoek near Stellenbosch in the become larger quite quickly, reaching peak or Western Cape, Cathedral Peak in the near peak reductions fairly early in the Drakensberg, Mokobulaan and Witklip on the rotation. Peak reductions under pine are Mpumalanga escarpment and Westfalia near reached around 15 years of age and at least Tzaneen. Additional data, although 5 years earlier under eucalypts. At the drier incomplete, from two small catchments at Ntabamhlope site flows ceased completely in Ntabamhlope near Estcourt in KwaZulu-Natal the fourth year after planting, which was also are also included. a dry hydrological year. It seems to be generally true that dry conditions will RESULTS accelerate the desiccation of the catchment after planting. For each successfully analysed experiment the estimated effects on total and low flows A Re-Analysis of the South African Catchment Afforestation Experimental Data Executive Summary, page ii Peak reductions per year (mean over 5 the highest flow reductions. This is probably consecutive years) range widely under pines because water demand is generally greater from 17 to 67 mm/10% and from 37 to 41 than supply in South African conditions mm/10% planted under eucalypts. The (situations where supply is unlimited are rare absolute reductions at Ntabamhlope have in this country). Good examples of this point only been measured for a few years but would are Cathedral Peak II and Tierkloof where 5- be much smaller than the figures for other year mean estimated peak reductions were eucalypt plantings. Peak absolute reductions 67 and 54 mm/10%, respectively. From this relative to expected flows occur at variable it also follows that wet years are those in stages within the rotation depending on site which the highest reductions are measured. specific conditions. Relative reductions, also This point is illustrated by Mokobulaan A & B over a five year window, have a narrower that were both dry for the latter half of the peak, from 6.6 to 10%/10% under pines to 9.8 rotations but because of differences in rainfall to 10%/l 0% planted under eucalypts. the estimated 5-year mean maxima were 41 and 17 mm/10% respectively. A new finding from this up-to-date analysis is that flow reductions are definitely diminished Conversely, low water availability can lead to towards the end of longer timber rotations, bigger relative reductions, earlier in the and this is true of both pines and at least one rotation, e.g. Mokobulaan B under pine eucalypt experiment. Obviously this trend is reached a 100% reduction in 12 years under clearest in the longer term experiments. The a dry cycle and the dry Ntabamhlope diminution of final flow reductions (mean over catchments, planted to a hardy eucalypt, last 5 years measured) compared to the reached 100% reductions in the fourth year of highest 5 year mean reductions ranges from the rotation. zero (no change over time, usually in short term experiments) to 60% and 50% less, for Fit of the CSIR empirical models absolute and relative measures respectively. The single eucalypt experiment in which this The empirical curves developed by the CSIR trend was observed was confounded by a are currently used to estimate the probable partial clearing (-10% cleared along the impact of afforestation schemes. The stream) priorto the restoration of streamflows. additional experiments analysed in this study However, the small area that was cleared is offer an opportunity to verify these models. It not likely to account for the large change in is clear from the study that the genera! flow reductions (48%). empirical models provide only an average and long-term estimate of the reductions caused DISCUSSION by afforestation. Age is far from being a complete predictor of streamflow reductions Understanding the drivers of afforestation as indicated by the large year by year effects variation in effects on flow.

Load more