The Use of Selected Species in Landscape Planning and Restoration of the Atlantic Forest, Brazil

40 41 4

THE USE OF SELECTED SPECIES IN LANDSCAPE PLANNING AND RESTORATION OF THE ATLANTIC FOREST, BRAZIL

Laury Cullen Jr., Alexandre Uezu, Cristiana Saddy Martins, and Claudio Benedito Valladares Padua

INTRODUCTION types in a region as a “coarse filter” approach to protecting biodiversity); and (3) evaluating the Successful and effective strategies for biodiversity requirements of selected focal species (Sanderson conservation must meet the needs of wildlife and et al. 2002; Coppolillo et al. 2004; Beier et al. 2011). people in biologically functioning and increasingly Because it is impossible to inventory every species human-dominated landscapes. Isolated protected in an ecosystem, scientists usually concentrate on a areas have had an increasing impact in protect- few key species. The use of selected species as a basis ing elements of biodiversity that promotes nature for site-based conservation has been widely used for conservation (Redford and Richter 1999; Redford designing landscape conservation. For example, et al. 2006; Leroux and Kerr 2013). Within these Lambeck (1997) presents a multi-species approach landscapes, the need for strictly protected areas for defining the attributes required to meet the with suitable habitat patches has been increasing needs of the biota in a landscape and the manage- emphasized for an array of species and components ment regimes that should be applied. His approach (Salazar et al. 2013). However, few systematic meth- builds on the concept of umbrella species, whose re- ods have been proposed to help in mapping and de- quirements are believed to encapsulate the needs of signing a conservation landscape. other species. It identifies a suite of “focal species,” In designing a protected area or a protected area each of which is used to define different spatial network (a regional system of protected areas), and compositional attributes that must be present conservationists generally use some combina- in a landscape and their appropriate management tion of three approaches: (1) mapping special ele- regimes. Miller et al. (1999) presented some con- ments (i.e., sites of high value such as wilderness cepts for using focal species in conservation action. areas, road-less areas, and location of rare species); They discuss the approach of focal species in plan- (2) seeking representation (i.e., including all habitat ning a reserve network and define focal species as

Aguirre040316ONATUS.indb 40 6/9/2016 5:34:08 PM 41

Restoration of the Brazilian Atlantic Forest 41

“organisms used in planning and managing nature habitat. It generated sufficient information to run reserves because their requirements for survival stochastic models that led to the creation of con- represent important factors in maintaining ecologi- servation scenarios for the species, combining wild cally healthy conditions.” A suite of landscape spe- population parameters with habitat availability and cies may be selected depending on resources, lead- captive bred animals, all of which were merged into ing to multiple, often overlapping focal species. a metapopulation management plan (Valladares- Based on this approach, a combination of three Padua et al. 2002). studies where birds, jaguars (Panthera onca), and Finally, it is demonstrated how the integration of the endangered black lion tamarin (Leontopithecus studies with species are helping to develop and to im- chrysopygus) are presented on how they can be used plement an agroforestry community-based program for defining suitable habitat patches for species con- to restore potential wildlife corridors, improving servation and site conservation planning in the frag- landscape management and protected area networks mented Atlantic forests of the interior, state of São in the Atlantic forest. Also, how a combination of Paulo, Brazil. For birds, analyses were conducted for species conservation and landscape restoration re- sensitive species. Based on the linear models gener- search can be used to create “informal” public poli- ated by the relation between the richness of sensi- cies that results in an appropriate regional landscape tive species and the landscape parameters from 1978 that is good both for biodiversity and humans. and 2003, past and recent ones (Uezu and Metzger 2007), the number of sensitive species were estimated in all forest patches in the study region, in- STUDY SITE: THE PONTAL DO dicating patches according to their priority of con- PARANAPANEMA servation, assisting to set a management plan for the conservation of biodiversity in the region. The Pontal do Paranapanema region is a wedge- For jaguars the “landscape detective” approach shaped region bounded on the north by the Parana was developed which is closely related to other focal River and on the south by the Paranapanema River, species focused conservation planning techniques forming the westernmost extremity of the state of such as umbrella, keystone, and flagship species São Paulo, Brazil (22°30′S, 52°20′W) (fig. 4.1). (Lambeck 1997; Caro and O’Doherty 1999; Miller Entirely forested, the region was decreed a et al. 1999; Sanderson et al. 2002). The landscape de- protected area in 1942 by the state of São Paulo, tective approach has certain special features. First, the “Grande Reserva do Pontal,” with about 2,600 it was developed for jaguar populations that still km2 (Leite 1998). Despite the protected status, the remain in highly fragmented habitats. Central in Pontal has suffered from numerous conflicts over this approach is to use the jaguar to detect core areas landownership, resulting in the widespread de- and isolated patches and stepping-stone habitats for struction of its forests for timber and cattle pasture landscape restoration and metapopulation conserva- during the past 50 years. In the mid-1990s, with pres- tion. Other approaches such as the umbrella species sure for land redistribution from the Landless Rural do not seem to relate specifically to enhancing con- Workers’ Movement (Movimento dos Trabalhadores nectivity, linkages, and defining specific lands and Sem Terra, MST) and other groups, many such areas for restoration. Also, the landscape detective properties were first occupied by MST affiliates. is more species-intrinsic in that the species-habitat These lands were later expropriated for public land relationship is the central point to this approach. In reform settlements, dramatically increasing the the landscape detective approach, species-habitat re- density of human occupation around the remaining quirements provided by habitat selection analysis are Atlantic forest fragments. used to build a habitat suitability function to define Today all that remains of the “Grande Reserva” is the landscape in which conservation must occur. the 36,000 ha Morro do Diabo State Park; the 6,000 The endangered black lion tamarin was also ha Black Lion Tamarin Ecological Station; and used as a focal species for landscape planning. about 16,000 ha of scattered patches which range The work with this critically endangered primate in size from one to 2,000 ha. The Morro do Diabo focused on the species biological studies and its State Park is well protected with legally demarcated,

Aguirre040316ONATUS.indb 41 6/9/2016 5:34:08 PM 42 43

42 Conserving Biodiversity

Brazil

State of São Paulo N

Road Rivers Forest

Pasture 22°10'0''S Agriculture 22°20'0''S 22°30'0''S

0510 20 30

km 22°40'0''S

52°50'0''W 52°40'0''W 52°30'0''W 52°20'0''W 52°10'0''W

Figure 4.1 The Pontal do Paranapanema and the remaining forest fragments. The large patch is the Morro do Diabo State Park Source: Uezu 2007.

nondisputed boundaries. It is the last significant of the region, and many of its endangered species, remnant of Atlantic forest in the west of São Paulo such as the jaguar and the black lion tamarin. state, where only 1.8% of the original natural vegeta- The Pontal do Paranapanema forests are con- tion remains. Because of this extensive loss of forest, sidered a transitional ecosystem, bordered by the conservation of the Morro do Diabo State Park, tropical evergreen broadleaf forest to the east, the Ecological Station, and other widely scattered, which originally covered the Atlantic coast- smaller, forest remnants is of utmost importance, as line, and the dry cerrado vegetation to the north they still harbor the rich and endemic biodiversity and west (Ab’Saber 1977). The Pontal region is

Aguirre040316ONATUS.indb 42 6/9/2016 5:34:09 PM 43

Restoration of the Brazilian Atlantic Forest 43

56°30’0”O 55°25’0”O 54°20’0”O 53°15’0”O 52°10’0”O

N BRAZIL W E 27

S 21°45’0”S

1 26 25 2 24 28 3 4 22°50’0”S

5 29 9 23°55’0”S 6 AY 8 7 GU 14 15 ARA P 13

10 25°0’0”S 30 11 17 20 16 12

19 21 18

22 26°5’0”S 23 64

31 27°10’0”S

ARGENTINA 02550100 km

Figure 4.2 Some important protected areas along the Upper and the Lower Parana River Ecoregion. Morro do Diabo State Park (25), Ivinhema State Park (24), Ilha Grande National Park (29), and Iguaçu National Park in Brazil and Argentina (30). Jaguar metapopulation dynamics were modeled within an area of approximately 50,000 km2 of the Upper Parana region, covering the range of 50 km on each side of the Parana and Paranapanema Rivers and from the town of Teodoro Sampaio (22˚31′36″ S, 52˚10′09″ W) to the town of Santa Terezinha de Itaipu (25˚21′36″ S, 54˚28′36″ W) in Brazil Source: Di Bitetti et al. 2003.

characterized by a pronounced dry season with For the jaguar study, the Upper Parana River annual precipitation averaging 1,370 mm, of which region was considered as the study area. This about 30% falls between April and September region strategically connects the Morro do Diabo (Valladares-Padua 1987, 1993). Most of the emer- State Park to other large protected areas along gent trees lose their leaves during the dry months the Parana River, including the Iguaçu National (Hueck 1972). The region is also known for its Park in Brazil and Argentina (fig. 4.2). These core generally nutrient-poor sand soils (Setzer 1949). areas form the basis of the Upper Parana Atlantic

Aguirre040316ONATUS.indb 43 6/9/2016 5:34:09 PM 44 45

44 Conserving Biodiversity

Rainforest Biodiversity Vision, a large-scale land- Number of sensitive birds = scape conservation plan developed by the World − 14.357 + 2.716 * []AREA + 0.4440 * []PROX. Wildlife Fund (Di Bitetti et al. 2003). Together, the Pontal and the Upper Parana River Because losses are more immediate and the time region are appropriate landscapes to use selected lag is less likely to occur in smaller remnants (Uezu species in landscape planning and restoration. 2007), to estimate the number of sensitive species in The ongoing destruction of the Atlantic forests is patches smaller than 300 ha, the model used was the confining remaining wildlife populations into net- one which comprise only the patch size (AREA) of works of small patches of suitable habitat. Thus, a 2003, the most recent year analyzed: reliable method of determining the requirements for effective landscape biodiversity conservation Number of sensitive birds = needs to be developed that includes metapopulation −13.073 + 4.4231 *A[]REA. and persistence of key species.

To verify the consistence of the multiple regression METHODOLOGY linear model, related to the period of 1978, we cor- related, using Pearson’s correlation, the observed Sensitive Birds as Focal Species richness of sensitive bird species in 21 forest patches In designing a bird conservation strategy, the rich- and in the control site, with the richness estimated ness of sensitive bird species was calculated in forest for the same set of patches using this model. patches based on a survey using the point count methodology (Blondel et al. 1970). We defined sen- Jaguars as Landscape Detectives sitive bird categories according to their responses to habitat fragmentation. It includes species which The spatial structure of the jaguar metapopula- occur only in the control areas or in large patches, or tion in the Upper Parana Region was based on those that decrease in abundance in smaller patches habitat data. This link between habitat data and the (Uezu 2007). Survey efforts were concentrated in metapopulation model was made possible by the 28 areas: 21 forest patches and 7 areas inside the Spatial Data program built in RAMAS GIS soft- control continuous forest. To ensure a wide range ware (Akçakaya 2005). The program used spatial of size, that represents the entire set of remnants, data on habitat requirements of the species, such which are relevant for bird conservation in the as GIS-generated maps of land cover and combin- region, seven large patches (400–1,500 ha), seven ing these data into a map of Habitat Suitability (HS) medium (100–200 ha), and seven small ones (30– with a habitat function. This map was then used 80 ha) were selected. Landscape parameters, patch to find habitat patches by identifying areas of high size, and degree of proximity (AREA and PROX, HS, where the jaguar population might exist and respectively), from three different years (1965, 1978, still survive (Cullen 2007). Topographic map layers and 2003) were used as explanatory variables to the and habitat categories mainly included those most variation of sensitive bird richness in nine differ- likely to explain jaguar habitat patches and popu- ent models (Uezu and Metzger 2007). Using AIC lation spatial distribution (table 4.1). These maps (Alkaike Information Coefficient), it was verified were classified from LandSat Images using super- that the model which combined patch size (AREA) vised classification (with ground truth) of the three and degree of isolation (PROX) of 1978 presented Landsat 7ETM satellite images that covered the the highest probability of being selected, present- Upper Parana River region. The analysis was done 2 with Erdas Imagine 8.4 and Arcview 3.3/Spatial ing a high coefficient of correlation (R adj. = 0.83). Therefore the parameters of this model were used Analyst. The term habitat was used to describe a to predict the richness of sensitive species in all layer of proportions of a habitat class as defined by large forest patches (> 300 ha) in the Pontal do vegetation type or other classifying factors used by, Paranapanema region. The model parameters or available to, an animal. Each habitat composition used are: summed to 100%. Errors in assigning an individual

Aguirre040316ONATUS.indb 44 6/9/2016 5:34:09 PM 45

Restoration of the Brazilian Atlantic Forest 45

Table 4.1 Habitat availability, use, and selection by jaguars (Panthera onca) in the Upper Parana Region, Brazil, 2007

Habitat Type Symbol (A) Availability (U) Use: Proportion of Ivlev’s Index of Proportion in the Study Jaguar Locations (%) Selectivity Site (%) Water water 6.997 4.104 −0.26057 Primary primfor 5.239 13.841 0.45087 forest Secondary secfor 2.605 5.153 0.32848 forest Alluvial aluv 0.970 1.025 0.02754 forest Dense dense- 4.977 10.045 0.33734 marshland marsh Open openmarsh 11.013 25.695 0.39995 marshland Agriculture agric 17.208 17.921 0.02030 Pasture pasture 50.991 22.216 −0.39307

radiolocation to a particular habitat class assumed conservative or a precautionary value, because that a jaguar used all habitats within a 100 m radius only a small portion of the landscape was assumed of a radiolocation in proportion to the area of that to be suitable. Second, neighborhood distance was habitat within the circle. In total, 11 adult jaguars used to identify nearby grid cells that belong to the were tracked with VHF and GPS collars from 2001 same patch (i.e., subpopulation) and may represent to 2005. the mean foraging distance of the species or the size of the home range. Based on the average home 2 2 Habitat Selection range area of 115.9 km ± 42.6 km the diameter of a circle shaped home range was calculated as 40 to 53 Habitat selection was determined as the dis- cells (1 cell = 300 m). The more conservative value tribution of all independent jaguar locations of 40 cells or 12 km was used. Based on the habi- in each habitat type in relation to the habitat tat map and these two parameters, RAMAS GIS availability during the study period. Habitat (Akçakaya 2005) identified the patch structure, selectivity was then defined by compar- which includes the size and location of habitat ing availability (A) and utilization (U), using patches. Each patch supports one subpopulation of Ivlev’s (1961) index of selectivity = (U – A)/ the metapopulation. This method of patch identi- (U + A). The link between the habitat map and fication has been previously described (Akçakaya the jaguar metapopulation was characterized by et al. 1995; Akçakaya 2000). This analysis resulted two parameters. First, a threshold habitat suitabil- in three populations, in which population one was ity HS was set as the minimum habitat suitability the Morro do Diabo population. value below which the habitat is not suitable for re- production and/or survival. Based on jaguar loca- Black Lion Tamarin Metapopulation tions along the entire Upper Parana basin, a value of 1.2 was used as the threshold HS (Cullen 2007). The black lion tamarin population occurred as The proportion of the study area with threshold a metapopulation dynamic in the study region HS at or above 1.2 is 9.1%. Thus, this represents a (Valladares-Padua 1993). To model the probability

Aguirre040316ONATUS.indb 45 6/9/2016 5:34:09 PM 46 47

46 Conserving Biodiversity

of extinction of the known black lion tamarin sub- RESULTS populations and the need for population manage- Sensitive Birds as Focal Species ment, simulations were run for each known subpopulation using Vortex software (Lacy 1993) and The estimated richness of sensitive birds presents a each of these subpopulations was considered an great variation throughout the forest remnants in the isolated subpopulation and modeled for a period region. Large patches (> 300 ha), which substantially of 100 years; each simulation was repeated 1,000 decreased in size from 1978 to 2003, have the highest times. In the same fashion a simulation was run for estimates of the richness of sensitive birds (fig. 4.3), the metapopulation (all subpopulations combined). whereas those large patches, which had already a sim- Both the probability of extinction and the expected ilar size in 1978 compared with 2003, had medium es- heterozygosity were tracked over time. When a timates of bird richness. For the medium-sized patch population size dropped to zero, it was counted (100–200 ha), a medium level of bird richness was also as an extinction event during that year. The prob- estimated, independently, whether they were larger ability of extinction was calculated to be the aver- in 1978 or not. Small patches (< 100 ha), presented the age extinction rate over all simulations. Expected lowest estimates of bird richness, regardless their size heterozygosity was calculated by monitoring allele in 1978. The observed and the estimated richness of frequencies over time in each subpopulation as well sensitive bird species in the sampled patches present a as in the metapopulation. The models were based high correlation (R Pearson = 0.918, fig. 4.4), support- mainly on data collected on the ecology and conser- ing that the selected model generated a good estimate vation biology of the black lion tamarins over seven of richness of sensitive birds, at least within the range years (Valladares-Padua 1993). of the sampled patch size.

Black lion tamarin presence Rivers Sensitive species estimation (n) 0 - 7 8 - 15 16 - 31

Figure 4.3 Estimation of the richness of sensitive bird species in forest patches in the Pontal do Paranapanema region, southeastern Brazil. Darker gray indicates priority areas for landscape management. The letters indicate the sampled patches: control area (Morro do Diabo State Park) (C), large patches (L), medium patches (M), and small patches (S) Source: Uezu 2007.

Aguirre040316ONATUS.indb 46 6/9/2016 5:34:10 PM 47

Restoration of the Brazilian Atlantic Forest 47

35 Based on these results, we defined (HS) as:

30 0.00203 *a[]gric +0.00275 *a[]luv es (n) 25 +0.03373 *d[]ensemarsh+0.040000*[]openmarsh 20 − 0.03931 *p[]asture +0.04509 *p[]rimfor sitive speci − 15 +0.03285 **s[]ecfor0.02606 *w[]ater ted sen 10

Esti ma The symbols within brackets refer to map layers 5 (table 4.1). All map layers were based on the same 0 vegetation map. Each cell of a layer gives the propor- 0 5101520253035 Observed sensitive species (n) tion of one habitat type in a grid cell of 300 by 300 m. This HS function determines the suitability of a location given various input maps describing envi- Figure 4.4 Significant relationship between the estimated and observed sensitive bird species in the surveyed ronmental variables. In other words, it attempts to patches, Pontal do Paranapanema, southeastern Brazil identify habitat patches (or subpopulations) from Source: Uezu 2007. the jaguars’ point of view. This function is used to calculate the HS for each location (cell) in the map. The model produced the habitat map (fig. 4.5) Jaguars as Landscape Detectives and a metapopulation structure with three subpopulations of suitable cells within the neighborhood Habitat selection was evaluated at gross scales that pro- distance of each other (fig. 4.6). These patches cov- vide a broad view of habitat requirements, and whether ered a total area of 4,105 km2 and had a total carry- jaguar use of habitat categories occurred in proportion ing capacity of 126 individuals (table 4.2). The patch to their availability in the study site (table 4.1). structure was realistic considering the remaining

W E

Habitat suitability –7 – –5 –4 – –2 –1 – 0 1 – 3 4 – 6

Figure 4.5 Habitat Suitability Map in the Upper Paraná region. Values for habitat suitability are represented in the scale below, with HS values ranging from −7.0 (least suitable) to 6.0 (most suitable). Jaguar metapopulation dynamics were modeled within an area of approximately 50,000 km2 of the Upper Paraná region, covering the range of 50 km on each side of the Paraná and Paranapanema Rivers

Aguirre040316ONATUS.indb 47 6/9/2016 5:34:10 PM 48 49

48 Conserving Biodiversity

48°0'0''W 47°0'0''W

02550km

Pop1

Pop2 23°0'0''S

N 24°0'0''S

Pop3 25°0'0''S

Figure 4.6 Patch structure of jaguar subpopulations identified by the model in the Upper Paraná region. Each gray tone indicates one habitat patch identified by the program. Patch 1 corresponds to the Morro do Diabo region in São Paulo state (Brazil). Patch 2 corresponds to the Ivinhema region in Mato Grosso do Sul state and Ilha Grande National Park region in Paraná state, both in Brazil. Patch 3 corresponds to jaguar subpopulations in eastern Paraguay along the Paraná River, and more specifically to the states of Canindeyú (Salto de Guairá) and Alto Paraná (Ciudade del Este), comprising an area with a series of protected areas such as Refugio Biológico Carapá, Reserva Natural Privada Itabo, Reserva Natural Privada Morombi, Reserva Biológica Mbaracayú, Reserva Biológica Pikyry, Refugio Biológico Tati Yupi, Reserva Biológica Itabo, and Reserva Biológica Limoy

habitat, known jaguar occurrences and the loca- numbers of jaguars are known to occur. This patch tion of some protected areas in the Upper Paraná- made up about 54% of the total area of all patches Paranapanema region. combined and had a carrying capacity of 64 indi- The largest patch for population 2 had an viduals. The three suitable patches covered only area of 2,224 km2 and comprised the Ivinhema 3.39% of the total landscape analyzed, a very small region and a large area to the south toward the area considering the area requirements of jaguars. Ilha Grande National Park, where considerable Patches have gaps between them, which represent

Aguirre040316ONATUS.indb 48 6/9/2016 5:34:10 PM 49

Restoration of the Brazilian Atlantic Forest 49

Table 4.2 Results of patch identification where total habitat suitability (HS) is the total value of the habitat suitability values (as described in table 4.1 and given in figure 4.5) in all cells included in a patch, Brazil, 2007

Patch Population Namea Total HS Area Area as Landscape % Carrying Number (km2) Patches % Capacity (K) 1 Morro do Diabo State Park 15,617 409 9.97 0.84 18 Region (São Paulo state) 2 Ivinhema State Park 55,422 2,224 54.57 1.34 64 Region (Mato Grosso do Sul state) and Ilha Grande National Park Region (Parana state) 3 Eastern Paraguay along the 38,404 1,472 35.86 1.22 44 Parana River, and more specifically to the states of Canindeyú (Salto del Guairá) and Alto Paraná (Ciudad del Este)b Total 109,443 4,105 100 3.39 126

a Including areas in the mosaic surrounding the current protected areas. b Comprising an area with a series of protected areas such as Refugio Biológico Carapá, Reserva Natural Privada Itabo, Reserva Natural Privada Morombi, Reserva Biológica Mbaracayú, Reserva Biológica Pikyry, Refugio Biológico Tati Yupi, Reserva Biológica Itabo and Reserva Biológica Limoy

unsuitable locations relative to jaguar habitat re- jaguar subpopulations. They can also contribute quirements and foraging distance. Major gaps to the design and restoration of an interconnected occur between the Morro do Diabo Region (popu- landscape. lation 1) and the Ivinhema State Park (population Each gray tone indicates one habitat patch iden- 2) and again, between Ivinhema and the southern tified by the program. Patch 1 corresponds to the populations identified in Eastern Paraguay along Morro do Diabo region in São Paulo state. Patch 2 the Paraná River (population 3). However, no corresponds to the Ivinhema region in Mato Grosso gaps were identified between the Ivinhema and do Sul state and Ilha Grande National Park region the Ilha Grande National Park. The landscape in Parana state. Patch 3 corresponds to jaguar between these two protected areas appeared suit- subpopulations in eastern Paraguay along the able for jaguars to establish their home ranges and Parana River, and more specifically to the states dispersal when the species habitat requirements of Canindeyu (Salto de Guaira) and Alto Parana and foraging distance were considered. Landscape (Ciudade del Este), comprising an area with a gaps identified between the three subpopula- series of protected areas such as Refugio Biológico tions might affect landscape connectivity and Carapá, Reserva Natural Privada Itabo, Reserva dispersal between them. The HS map has great Natural Privada Morombi, Reserva Biológica potential to be used to identify stepping-stone Mbaracayu, Reserva Biológica Pikyry, Refugio corridors. These are cells with high suitability Biológico Tati Yupi, Reserv Biológica Itabo, and value outside the patches identified and can link Reserva Biológica Limoy.

Aguirre040316ONATUS.indb 49 6/9/2016 5:34:10 PM 50 51

50 Conserving Biodiversity

Black Lion Tamarin Metapopulation DISCUSSION Conservation Sensitive Birds as Focal Species

The model used simple life-table calculations to The knowledge about the past landscape structure estimate the population growth rates and genera- was useful to evaluate more precisely how the bio- tion lengths that are be used in the simulations. diversity is distributed throughout the studied land- Extinction rates in all modeled subpopulations scape, especially regarding the sensitive species. The failed to meet the minimum tolerance criteria of analysis of the past landscape parameters also dem- 2% probability of extinction in 100 years (table 4.3). onstrated that the largest patches cannot hold the When isolated, the four smaller populations (those current diversity in the long term, resulting in a high outside of Morro do Diabo State Park) have low or number of species likely to become extinct in the negative growth rates, high probability of extinction near future (Tilman et al. 1994). The results of this and become highly inbred. Corridors were modeled work also indicate that to avoid these future losses using dispersal among populations for tamarins of it is necessary to increase the forested area available both sexes from three to six years, with 50% mortal- for those sensitive species. This can be done carry- ity occurring during dispersal. The results of this ing out two different types of landscape manage- modeling shows that at the level of connectivity, the ment: increasing the size of the large forest patches creation of corridors has a significant effect on the to a similar extension of those in 1978 or implement- viability of the four smaller populations, with little ing corridors among these areas, increasing the land- negative impact upon black lion tamarins inhabit- scape connectivity and giving the species access to ing the Morro do Diabo Park. This means that all several patches. The logistical and financial factors four local populations have essentially no risk of ex- favor the second option. For the first option to be ef- tinction when they have a chance to be integrated, fective very large areas should be restored to comple- and they are able to maintain population sizes near ment the size of the larger remnants. carrying capacity with levels of gene diversity above In order to recover all the forested area lost in 90% (Valladares-Padua 1993). the last 25 years regarding the six largest patches

Table 4.3 Simulation results for black lion tamarin (Leontopithecus chrysopygus) scenarios over 100 years, São Paulo, Brazil, 1993

Population Scenario PE N GD MTE Metapopulation Isolated 0.000 1,090 0.986 Corridors 0.000 1,116 0.985 Morro do Diabo Isolated 0.000 1,067 0.985 Corridors 0.000 1,004 0.985 Tucano Isolated 0.446 14 0.608 68 Corridors 0.000 45 0.952 Ponte Branca Isolated 0.544 9 0.569 62 Corridors 0.000 38 0.934 Santa Maria Isolated 1.000 — 0.914 24 Corridors 0.000 15 — Santa Monica Isolated 1.000 — 0.901 22 Corridors 0.006 14 — 56

Note: PE = probability of extinction; N = mean population size; GD = gene diversity; MTE = mean time to extinction in years.

Aguirre040316ONATUS.indb 50 6/9/2016 5:34:10 PM 51

Restoration of the Brazilian Atlantic Forest 51

in the region, it would be necessary to implement and habitat enhancement and restoration (Eizirik more than 7,000 ha of forest, and approximately the et al. 2002; Cullen et al. 2005). To evaluate the ef- same extension these remnants cover today. On the fectiveness of such a diverse set of management other hand, with a smaller effort, the implantation options requires a method that can integrate dif- of corridors connecting the larger patches would ferent types of information into a single assessment increase many times the forested area available for framework. It also needs to incorporate a variety of the sensitive species, especially if these connections factors and the interactions for a population under encompass the Morro do Diabo State Park, with several threats, and a fragmented distribution in the 37,000 ha (fig. 4.2). However for the corridors to be landscape. One of the most important strengths of functional for these particular species, they must the landscape detective model is the integration of be wide enough to include interior forest habitat, information and in particular the integration of the as many of them are edge sensitive. Alternatively, metapopulation models. the matrix surrounding the corridors could be In the Upper Parana Region, the time is criti- managed, changed by a less resistant culture, what cal for protecting land areas under a plan that will could favor the landscape connectivity and decrease ensure habitat contiguity with future land develop- the edge effect along the corridor (Mesquita et al. ment. The jaguar in the Upper Parana ecosystem 1999). Studies showed that the efficiency of stepping represents an endangered population and isolated stones and corridors are regulated by the matrix from the Pantanal and the Coastal Atlantic forests. permeability and are species-dependent (Baum The distribution of jaguars reflects relative habitat et al. 2004). This study highlights the importance conditions along a human-impacted landscape of im- of using information about the history of the land- portance to wildlife conservation. This study identi- scape modification to set priority areas for conser- fied three large suitable patches for jaguar conserva- vation. If no management is conducted to revert, tion, which together were about 4,100 km2 in area or at least partially, the fragmentation process in the equivalent to 8% of the potential habitat in the study region, several losses might occur, as it happened area. These large patches represent core areas for jag- in other Atlantic forest regions (Christiensen and uars in the region and they should be proposed for Pitter 1997; Ribon et al. 2003). protected-area management, which may include a combination of different land uses such as intensive Jaguars as Landscape Detectives use areas, buffer zones or intermediate use areas, and some strictly protected areas that could be linked by The threats, ecology and distribution, and manage- wildlife corridors and other suitable areas identified ment options of the jaguars in the Upper Parana by this model. The suitability map also helped in River Corridor necessitate the use of models to identifying landscape gaps or strategic transit ref- evaluate options for their conservation and man- uges or stepping stones for dispersing jaguars that agement. The results of spatial distribution of jag- could improve the dispersal potential of corridors. uars in this region, and their dispersal ecology and This information will lead to recommendations for demography suggested a metapopulation struc- corridor restoration in a human-dominated land- ture, with distinct but interacting local subpopula- scape for jaguar conservation (Cullen 2007). tions inhabiting relatively suitable habitat patches separated by less suitable areas. The jaguars in this Black Lion Tamarin Metapopulation region are threatened by several factors, including Conservation habitat loss, habitat fragmentation, road mortality, and mortality resulting from their interaction with The results of the Vortex model suggest that if livestock (Crawshaw et al. 2004; Cullen et al. 2005, treated individually, black lion tamarin subpopu- 2013). Each of these factors affects a different aspect lations, with the exception of the Morro do Diabo, of the jaguar metapopulation. There are also several have more than 50% chances of going extinct in the possible types of management actions that may ben- next 100 years. Even the Morro do Diabo subpopu- efit these populations, including habitat protection, lation, with its comparatively large size, does not increasing connectivity, decreasing road mortality, meet the survival criteria established for the species.

Aguirre040316ONATUS.indb 51 6/9/2016 5:34:11 PM 52 53

52 Conserving Biodiversity

These results call for immediate action to be implemented if it did not become a public policy. manage the subpopulations in order to obtain a The solution came with the help of the state and fed- minimum viable size. Their ability to survive in eral Public Attorney’s offices in the region (Cullen different habitats, their recently discovered sub- et al. 2005), as these agencies have been supporting populations, and their ecological flexibility leads the overall idea of planning where efforts are to be to the conclusion that black lion tamarins will made for conservation and forest restoration. This probably respond well to metapopulation man- has been of critical importance, as in Southeastern agement procedures. This is, of course, a major Brazil it is mandatory by law to conserve 20% of shift from the total extinction prediction during all rural properties (called legal reserve). The law, the 1970s (Coimbra-Filho 1970) or from the esti- however, is not strictly enforced and as a result mated 100 individuals on the verge of extinction many properties lack natural coverage. To become in the wild during the 1980s (Coimbra-Filho and legal, landowners need to replant legal reserve in Mittermeier 1977). While the modeling results their properties, to either tell the authorities where are pretty straightforward, the conclusions herein in their properties the forest reserve is located or are based on a comprehensive interpretation of where it will be restored. Traditionally, land author- the results. The model is simple, the data are the ities accept the reserve position proposed by the best available though sometimes scarce, and the landowner, but in the Pontal do Paranapanema this involved biological and ecological processes are so is no longer the case. A “consented public policy” is complex that there is no way to rely completely on in place due to a memorandum of understanding the extinction probabilities of black lion tamarins. that was signed in 2005 by the majority of the deci- But despite the comprehensive interpretation of sion makers in the region. As an addendum to the the results and the many assumptions in modeling, memorandum, a restoration map was created using they certainly show that a managed total metapop- the legal reserves and gallery forests to form forest ulation is more secure than the subpopulations in corridors reconnecting all forest fragments in the isolation. region. In the last 20 years, IPÊ’s landscape conservation program has demonstrated high rates of INFLUENCING POLICY AND success in working with small landowners in the LANDSCAPE MANAGEMENT development of agroforestry buffer zones, corridors, and stepping-stones, and in influencing In 1997, after more than 10 years of conservation policy through the use of scientific studies. These research in this region, IPÊ’s scientists could un- studies are the basis for a conservation plan for derstand the importance of landscape planning the region that aims at better protecting not only for effective conservation of biodiversity and as the forest fragments but will also contributing an essential way to save species and their habi- to species survival through the reforestation of tats. Research results had suggested the need of corridors vital for species movement (Uezu and protecting the remaining forest fragments, the Cullen 2012). implementation of biological corridors, and the Policy recommendations have been based on creation of buffer zones for the fragments. The use the biological research, and in many instances, of agroforestry to create buffer zones to protect the measures have been adopted by the government surviving forest fragments and forest restoration (Valladares-Padua et al. 2002). For example, IPÊ to connect them in a landscape forest network is successfully lobbied for the creation of a buffer rescuing, in some ways, the Great Reserve of the zone around forest fragments when new settlers Pontal (Valladares-Padua et al. 1997), the original are given land, corridors, and legal reserves. The protected area of the 1940s. scientists at IPÊ have also influenced the creation Although the plan has the support of many of a 6,600 ha new protected area, the Black Lion regional stakeholders, including leaders of the Tamarin Ecological Station, based on biological Landless Movement, government officials, and surveys of the region. Also, a main forest cor- even some large landowners, it would be difficult to ridor is been established to connect the Morro

Aguirre040316ONATUS.indb 52 6/9/2016 5:34:11 PM 53

Restoration of the Brazilian Atlantic Forest 53

Rivers Legal reserves proposal Permanent protected areas Urban areas 0510 20 km Farms Small properties Forest remnants

Figure 4.7 The light gray bands in figure 4.7 indicate the landscape connectivity planning with agroforestry and forest corridors areas established. Dark gray are forest fragments and protected areas remaining. White parcels are private- owned properties. In the center, the Morro do Diabo State Park

do Diabo State Park and the Ecological Station ACKNOWLEDGMENTS (fig. 4.7). Ultimately, the fate of biodiversity in the Pontal The research presented was funded by grants from do Paranapanema and the Atlantic forest within The Liz Claiborne Art Ortenberg Foundation, relatively small forest fragments, isolated parks, Wildlife Trust, The Conservation, Food and and reserves remain in the hands of the people Health Foundation, FAPESP (The State of São that live around these areas and coexist with these Paulo Research Foundation, process no. 02/01746- species. However, it is the responsibility of man- 1). Restoration programs underway are currently aging agencies to resolve local conflicts that are funded by the BNDES Brazil, Natura, Duke Energy, inevitable in the interface between the natural and Cesp, Funbio, The Whitley Fund for Nature, and the human-modified worlds. Only through the in- The Center for Economy, Environment and Society tegration of applied research, implementation of at Columbia University, New York. management recommendations derived from these findings, involvement of NGOs and universities, co-management of protected areas, participation REFERENCES of local people through the community-based landscape restoration program, agroforestry extension Ab’Saber, A.N. 1977. Os domínios morfoclimáticos da and environmental education programs, and ap- América do Sul. Geomorfologia 52:1–23. propriate policies will the species have a chance to Akçakaya, H.R. 2000. Viability analyses with habitat- survive. based metapopulation models. Pop Ecol 42:45–53.

Aguirre040316ONATUS.indb 53 6/9/2016 5:34:11 PM 54 55

54 Conserving Biodiversity

Akçakaya, H.R. 2005. RAMAS GIS: Linking spatial Cullen, L., Jr., D.A. Sana, F. Lima, K.C. Abreu, and data with population viability analysis. Version 5.0. A. Uezu. 2013. Selection of habitat by the jaguar, Applied Biomathematics, Setauket, NY. Panthera Onca (Carnivora: Felidae) in the Upper Akçakaya, H.R., M.A. McCarthy, and J. Pearce. 1995. Paraná River. Zoologia 30:379–87. Linking landscape data with population viability Di Bitetti, M.S., G. Placci, and L.A. Dietz. 2003. A bio- analysis: Management options for the helmeted diversity vision for the Upper Paraná Atlantic honeyeater. Biol Conserv 73:169–76. Forest Eco-region: Designing a biodiversity Baum, K.A., K.J. Haynes, F.P. Dillemuth, and J.T. conservation landscape and setting priorities Cronin. 2004. The matrix enhances the effective- for conservation action. World Wildlife Fund, ness of corridors and stepping stones. Ecology Washington, DC. 85:2671–2676. Eizirik, E., C.B. Indrusiak, and W.E. Johnson. 2002. Beier, P., W. Spencer, R.F. Baldwin, and B.H. McRae. Análisis de la viabilidad de las poblaciones de 2011. Towards best practice for developing regional jaguar: Evaluación de parámetros y estudios de connectivity maps. Conserv Biol 25: 879–92. caso en tres poblaciones remanentes del sur de Blondel, J., C. Chessel Ferry, and B. Frochot. 1970. sudamérica. In R.A. Medellín, C. Equihua, C.B. La méthode des indices ponctuels d´abondance Chetkiewicz, P.G. Crawshaw Jr., A. Rabinowitz, (I.P.A.) ou des relevés d´avifaune par “stations K.H. Redford, J.G. Robinson, E.W. Sanderson, d´écoute.” Alauda 38:55–71. and A.B. Taber (eds.) El jaguar en el nuevo mi- Caro, T.M., and G. O’Doherty. 1999. On the use of sur- lenio, pp. 403–11. Fondo de Cultura Economica, rogate species in conservation biology. Conserv Universidad Nacional Autonoma de Mexico, and Biol 13:805–14. the Wildlife Conservation Society, Mexico City, Christiensen, M.M., and E. Pitter. 1997. Species loss Mexico. in a forest bird community near Lagoa Santa in Hueck, K. 1972. As florestas da América do Sul. Editora Southeastern Brazil. Biol Conserv 80:23–32. Polígono e Editora Universidade de Brasilia, São Coimbra-Filho, A. F. 1970. Considerações gerais e Paulo, Brasilia, Brasil. situação atual dos micos-leões escuros Leontideus Ivlev, V.S. 1961. Experimental ecology of the feeding of chrysomelas (Kuhl, 1820) e Leontideus chrysopygus fishes. Yale University Press, New Haven, CT. (Mikan, 1823). Rev Bras Biol 30:249–68. Lacy, R.C. 1993. VORTEX: A computer simulation Coimbra-Filho, A.F. and R.A. Mittermeier. 1977. model for population viability analysis. Wildl Res Conservation of the Brazilian lion tamarins 20:45–65. Leontopithecus rosalia. In P. Rainier and G.H. Lambeck, R.J. 1997. Focal species: A multi-species Bourne (eds.) Primate conservation, pp. 123–34. umbrella for nature conservation. Conserv Biol Oxford University Press, New York. 11:849–56. Coppolillo, P., H. Gomez, F. Maisels, and R. Wallace. Leite, J.F. 1998. A ocupação do Pontal do Paranapanema. 2004. Selection criteria for suites of landscape Editora Hucitec, Fundação UNESP, São Paulo, species as a basis for site-based conservation. Biol Brasil. Conserv 115:419–30. Leroux, S.J., and J.T. Kerr. 2013. Land development in Crawshaw P.G., J.K. Mahler, C. Indrusiak, S.M.C. and around protected areas at the wilderness fron- Cavalcanti, M.R.P.L. Leite, and K.M. Silvius. 2004. tier. Conserv Biol 27:166–76. Ecology and conservation of the jaguar (Panthera Mesquita, R.C.G., P. Delamonica, and W.F. Laurance. onca) in Iguacú National Park. In K.M. Silvius, 1999. Effects of surrounding vegetation on edge- R. Bodmer, and J.M.V. Fragoso (eds.) People in related tree mortality in Amazonian forest frag- nature: Wildlife conservation in South and Central ments. Biol Conserv 91:129–34. America, pp. 271–85. Columbia University Press, Miller, B., R. Reading, J. Strittholt, C. Carroll, R. Noss, New York. M. Soulé, O. Sanches, J. Terborgh, D. Brightsmith, Cullen, L., Jr. 2007. Jaguars as landscape detectives T. Cheeseman, and D. Foreman. 1999. Using focal for the conservation of the Atlantic forest, species in the design of nature reserve networks. Brazil. Ph.D. Dissertation. University of Kent, Wild Earth Winter 4:81–92. Canterbury, UK. Redford, K.H., and B.D. Richter. 1999. Conservation Cullen, L., Jr., K.C. Abreu, D. Sana, and A.F.D. Nava. of biodiversity in a world of use. Conserv Biol 2005. Jaguars as landscape detectives for the 13:1246–56. Upper Paraná River corridor, Brazil. Natureza e Redford, K.H., J.G. Robinson, and W.M. Adams. 2006. Conservação 3:43–58. Parks as Shibboleths. Conserv Biol 20:1–2.

Aguirre040316ONATUS.indb 54 6/9/2016 5:34:11 PM 55

Restoration of the Brazilian Atlantic Forest 55

Ribon, R., J.E. Simon, and G.T. De Mattos. 2003. Bird conservação. In A. Peace, A. Uezu, M.L. Lorini, extinctions in Atlantic forest fragments of Viçosa and A. Cunha (eds.) Conservação da biodivers- Region, Southeastern Brazil. Conserv Biol 17:1827–39. idade com SIG, pp. 13–23. Oficina de Textos, São Salazar, C.L., C.D.L. Orme, P.C. Rasmussen, T.M. Paulo, Brasil. Blackburn, and K.L. Gaston. 2013. The perfor- Uezu, A., and J.P. Metzger. 2007. Landscape dynamic mance of the global protected area system in affecting bird diversity in an Atlantic forest capturing vertebrate geographic ranges. Biodivers fragmented region: An evidence for time lag Conserv 22:1033–47. response. In R.G.H. Bunce, R.H.G. Jongman, Sanderson, E.W., K.H. Redford, A. Vedder, P.B. L. Hojas, and S. Weel (eds.) 25 years landscape Coppolillo, and S.E. Ward. 2002. A conceptual ecology: Scientific principles and practice. model for conservation planning based on land- Proceedings of the 7th IALE World Congress 8–12 scape species requirements. Landscape Urban Plan July, pp. 386-387. Wageningen, The Netherlands, 58:41–56. IALE Publication series 4. IALE:http://www. Setzer, J. 1949. Os solos do Estado de São Paulo. landscape-ecology.org Conselho Nacional de Geografia, Rio de Janeiro, Valladares-Padua, C. 1987. Black lion tamarin Brasil. (Leontopithecus chrysopygus): Status and conserva- Tilman, D., R.M. May, C.L. Lehman, and M.A. Nowak. tion. MSc Thesis. University of Florida. 1994. Habitat destruction and the extinction debt. Valladares-Padua, C. 1993. The ecology, behavior Nature 371:65–66. and conservation of the black lion tamarins Uezu, A. 2007. Composição e estrutura da comunidade (Leontopithecus chrysopygus, MIKAN, 1823). Ph.D. de aves na paisagem fragmentada do Pontal do Dissertation. University of Florida. Paranapanema. Ph.D. Dissertation. University of Valladares-Padua, C., S.M. Padua, and L. Cullen. 2002. São Paulo, Brasil. Within and surrounding the Morro do Diabo State Uezu, A., and L. Cullen. 2012. Da fragmentação Park: Biological value, conflicts, mitigation and Florestal à restauração da paisagem: Aliando con- sustainable development alternatives. Environ Sci hecimento científico e oportunidades legais de Policy 5:69–78.

Aguirre040316ONATUS.indb 55 6/9/2016 5:34:11 PM