Influence of the landscape matrix on the abundance of arboreal primates in fragmented landscapes

Please download to get full document.

View again

of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Pets & Animals

Published:

Views: 42 | Pages: 9

Extension: PDF | Download: 0

Share
Description
Composition of the landscape matrix of surrounding forest fragments is thought to be critically important to the survival of arboreal primates because it offers structures that help the animals move between fragments and other foraging sites.
Tags
Transcript
  ORIGINAL ARTICLE Influence of the landscape matrix on the abundance of arborealprimates in fragmented landscapes Gilberto Pozo-Montuy  • Juan Carlos Serio-Silva  • Yadira M. Bonilla-Sa´nchez Received: 7 September 2010/Accepted: 25 November 2010/Published online: 6 January 2011   Japan Monkey Centre and Springer 2010 Abstract  Composition of the landscape matrix of sur-rounding forest fragments is thought to be criticallyimportant to the survival of arboreal primates because itoffers structures that help the animals move betweenfragments and other foraging sites. However, little isknown about the composition of the matrix used by theseanimals. The aim of this study was to quantitatively assessthe importance of the landscape matrix and its effects onprimate abundance, using black howler monkeys (  Alouatta pigra ) living in a landscape fragmented by the expansion of agriculture and pastures for livestock in southeasternMexico. In 2008, a complete census of the monkeys wascarried out across the 2000-ha landscape matrix, and forevery site where we observed monkeys, we recorded can-opy height, tree basal area, food-source abundance, anddistance to the nearest fragment. A total of 244 howlermonkeys, distributed among 48 groups (including six sol-itary males) were counted in the matrix. Mean troop sizewas 5.6  ±  2.8 individuals, and the mode was three indi-viduals. The highest number of troops and greatest howlermonkey abundance were recorded in the isolated trees, theeucalyptus plantation, and orchards. A generalized linearmodel revealed that monkey abundance tended to be higherin matrix elements with higher canopy height, greater foodavailability, and closest to rainforest fragments. Theseresults suggest that it is necessary to take into account themany elements of the landscape when drawing up con-servation and habitat management plans, particularly inorder to establish connectivity among the fragments andelements of the matrix with native trees. Keywords  Fragmentation    Isolated trees   Living fences    Landscape matrix    Arboreal primates   Habitat management    Conservation Introduction The high rates of tropical deforestation, concurrent withhuman land-use expansion, have given rise to the loss andfragmentation of forest habitats (Burgess and Sharpe 1981;Lord and Norton 1990; Achard et al. 2002). Fragmentation drastically alters habitat availability and quality for wildanimals (Saunders et al. 1991; Bierregaard et al. 1992; Li 2004) and results in the creation of landscapes comprisedof vegetation remnants of varying sizes and degrees of disturbance, and isolation, surrounded by a matrix that ismainly agricultural in nature and occasionally has arborealelements such as living fences, forest plantations, andisolated trees (Lord and Norton 1990; Laurance et al. 1998). Habitat fragmentation—frequently associated with adecrease in animal population size and increase in their risk of extinction—is considered one of the greatest threats totropical biodiversity (Laurance and Bierregaard 1997;Chapman et al. 2007), although the seriousness of thethreat depends on the specific characteristics of each pop-ulation, species, site, or region (Johns 1986; Offerman et al.1995). G. Pozo-Montuy ( & )    Y. M. Bonilla-Sa´nchezPosgrado en Neuroetologı´a, Instituto de Neuroetologı´a,Universidad Veracruzana, Avenida Luis Castelazo Ayala s\nCol. Industrial Las Animas, 91190 Xalapa, Veracruz, Mexicoe-mail: gmontuy@gmail.comG. Pozo-Montuy    J. C. Serio-Silva    Y. M. Bonilla-Sa´nchezEstacio´n de Primatologı´a y Vida Silvestre A.C., Balanca´n,Tabasco, MexicoJ. C. Serio-SilvaRed de Biologı´a y Conservacio´n de Vertebrados,Instituto de Ecologı´a AC, Xalapa, Veracruz, Mexico  1 3 Primates (2011) 52:139–147DOI 10.1007/s10329-010-0231-5  Arboreal primates are among the species that are mostsensitive to habitat fragmentation. These monkeys needtrees with specific characteristics to meet their needs forfood, refuge, resting sites, and reproduction (Stevenson2001; Mendes-Pontes and Soares 2005; Arroyo-Rodrı ´guezet al. 2007) and thus are very sensitive to changes invegetation structure within their range (Balko and Under-wood 2005; Michalski and Peres 2005). In rainforest fragments, there can be a decrease in the availability of treespecies that are important to arboreal monkeys (Arroyo-Rodrı´guez and Mandujano 2006; Chapman et al. 2010) because habitat reduction and edge effects typically causemortality and a decrease in mean tree size (Saunders et al.1991). This makes it necessary for arboreal monkeys tomove across the matrix to use different fragments orarboreal elements in their search for resources (Andersonet al. 2007; Asensio et al. 2009). Laurance (1991) reported that the mammal species that make frequent use of thelandscape matrix typically have stable numbers or increasetheir population size, whereas in contrast, species thatavoid going out into the matrix decrease in numbers ordisappear. This means that in a scenario such as that of thefragmented landscape described, the matrix may well be acritical factor in animal survival, though to our knowledge,the importance of the matrix has never been formallystudied in detail for arboreal primates.To date, most studies on arboreal primates and the otherspecies that inhabit fragmented landscapes have beenconfined to the interior of the fragments and have virtuallyignored the effect of the matrix (Gascon et al. 1999; Chapman and Peres 2001; Ricketts 2001). Although it has been inferred from anecdotal reports that the matrix isimportant to the survival of arboreal primates (Carpenter1934; Scho¨n-Ybarra 1984; Bicca-Marques and Calegaro-Marques 1995; Serio-Silva and Rico-Gray 2000; Harris and Chapman 2007), only recently have there been studies onsome aspects of the use of the matrix by primates, withreports that the presence of crops and other types of arboreal cover offer alternate food sources for guenons,howlers, siamangs, lemurs, and colobus monkeys (Ganz-horn 1987; Michon and de Foresta 1995; Naughton-Treves 1998; Chapman and Onderdonk  1998; Anderson et al. 2007; Pozo-Montuy and Serio-Silva 2007; Asensio et al. 2009). On the other hand, it has been reported that thefrequency of the movement of monkeys across the matrixtends to increase as the distance between fragmentsdecreases or as the number of vegetation corridors betweenfragments increases (Estrada and Coates-Estrada 1996;Swart and Lawes 1996; Mandujano and Estrada 2005; Anzures-Dadda and Manson 2007).Based on the above, we can say that an arboreal primatethat lives in a fragmented landscape and is capable of usingthe matrix for moving and feeding likely has high long-term potential viability in fragmented landscapes. How-ever, it is not known which of the matrix attributes—i.e.,the availability of food species, the distance to the edge of the nearest fragment, or the physical structure of the veg-etation (the height and basal area of the trees in thematrix)—can help or hinder the abundance of some arbo-real primate species. With this study, we aim to addressthese points by answering the central question: Whichcharacteristics of the matrix help arboreal monkeys tosurvive in fragmented landscapes? We use an arborealprimate (  Alouatta pigra ) living in a fragmented region insoutheastern Mexico to answer this question.  A. pigra  ismainly tree dwelling and is most frequently found in theupper canopy of the rainforest. Diet consists of leaves andseasonally available fruit (Pavelka and Knopff  2004; Pozo-Montuy and Serio-Silva 2006). This species was consid-ered suitable for addressing our research questions as itscharacteristics (diurnal activity, slow movements, cohesivegroup, and tolerance of human presence) make it easier toconduct a full census. Specifically, we address the fol-lowing: What is the size of the howler monkey populationin the landscape matrix? Where in the matrix is  A. pigra most abundant? What does  A. pigra  do in the differentelements of the matrix? Which attributes of the matrixelements affect  A. pigra  abundance? Methods Study areaThe study area is located at 91  30 0 N lat., 17  40 0 W long.,and covers 2000 ha in the municipality of Balanca´n in thestate of Tabasco in southeastern Mexico and is bounded byseveral barriers that are both natural (mainly the Usumac-inta River and several lagoons) and anthropogenic (roadsand wire fences, among others; Fig. 1). It forms part of theLower Basin of the Usumacinta River, a region charac-terized by vast plains that flood during the rainy season(August–October). Mean annual precipitation is 1595 mm,and mean annual temperature is 27.9  C (CNA 2008).The srcinal vegetation in the region was described asrainforest, semideciduous tropical forest, and spiny ever-green bloodwood (  Haematoxylum campechianum ) forest(Lo´pez-Mendoza 1980). The government’s developmentplans during the 1960s resulted in increased colonization of the land and its conversion for pasturing livestock andgrowing crops, all of which caused the loss and fragmen-tation of the 115000 ha of srcinal forest (Reyes-Castillo1978; Pozo-Montuy et al. 2008). The conversion of forests to pasture has resulted in a loss of moisture and has createdan unusual climate, with long droughts (8 months) lastingfrom February to September, when historically the dry 140 Primates (2011) 52:139–147  1 3  season was only 3 months (from March to May) (PDM2005). Climate change has had another significant impact,as very dry years alternate with extensive flooding andincreased silting in rivers, streams, and lakes, which occursin the absence of forests that would normally retain the soilaround bodies of water (Chapman and Chapman 2003).Currently, the landscape comprises vegetation fragmentssurrounded by a matrix that is primarily pasture, along withother elements, such as living fences and isolated trees, aswell as areas used for silviculture and growing fruit (Fig. 2;Pozo-Montuy 2006). Silviculture in the area includesexotic species, mainly eucalyptus ( Eucalyptus  spp.), teak ( Tectonia grandis ), and melina ( Gmelina arborea ), withsome plantations of cedar ( Cedrela odorata ). In Balanca´n,about 12000 ha are devoted to silviculture.Monkey census in the landscape matrixTo quantify  A. pigra ’s use of the different matrix elementsby estimating its abundance, three expert observers, on footor on horseback, evaluated how many monkeys were usingeach element of the landscape matrix in the area. Thiscensus was done 17 days/month from March to December2008 (10 months and 1,369 person hours). These dailysurveys were done at a mean speed of 8  ±  2 km/day,between 0700 and 1600 hours. The time required to check for the presence of howlers in each element varieddepending on the size of the element (mean 40 min/ ha  ±  12 min). Each time visual contact was establishedwith a troop or individual, we recorded the matrix element,the age and sex of each individual, the number of indi-viduals per troop, and their activities when sighted. Agecategories were adult (an independent individual with awell-developed hyoid and chin, and [ 3 years old), juvenile(an individual that moves independently of its mother, isvery active, and [ 1 year old), and infant (an individual thatis totally dependent on its mother) (Rosales-Meda et al.2007). The term immature refers to juveniles and infants.Each element of the matrix was checked once during of thestudy and, in order to avoid recording the same troop morethan once, we used a GPS to record coordinates of each siteand measure the distance to the nearest group, as well asusing the sex–age composition and the physical charac-teristics of the monkeys (size, facial scars, scars on limbs,and fur color around the face, on the scrotum and vulva) asspecific identifiers for each troop. Additionally, data from aconcurrent radiotelemetry project allowed us to determinethe ranging distances for the howlers and thus reduce thesearch effort required in each matrix. Most of the matrixelements are less than half a hectare in size, which helpedus to do complete censuses. There was only one eucalyptusplantation (200 ha) in the area, which took 2 days to cen-sus. Each troop was observed with binoculars for at least30 min, and the census was considered complete when allthree observers had agreed on the composition of the troop.The probability of individuals from a troop or solitaryindividuals not being detected during our survey wasminimal, as three observers thoroughly inspected all the Fig. 1  Location of the municipality of Balanca´n in the state of Tabasco in southeastern Mexico Fig. 2  Examples of   a  natural barriers that circumscribe the study area(Usumacinta River),  b  pasture lands \ 1-m tall with some livingfences, roads, and isolated trees. Picture from Leo´n D. Olivera-Go´mezPrimates (2011) 52:139–147 141  1 3  elements of the matrix. Furthermore, the characteristics of the species make it easy to spot them in disturbed sitesbecause the monkeys do not show any fusion to subgroupsand are not hunted. They do not exhibit cryptic behavior,and they vocalize throughout the day. Because of this, thereis negligible variability in the detectability of the howlersbetween habitats.Attributes of the landscape matrixWe define habitat fragments as the native vegetation(excluding living fences) with a continuous canopy C 0.2 ha in area (this size threshold was selected based onthe size of the smallest inhabited fragment) and C 10 m tall(Mandujano et al. 2006; Anzures-Dadda and Manson 2007; Zunino et al. 2007). All types of plant cover not meetingthese criteria were considered matrix elements. There were14 different types of matrix elements (Table 1), and theirclassification was based on previous studies (Grunblattet al. 1989; Anderson et al. 2007). The characteristics of each matrix element that wereconsidered likely to affect the abundance of monkeyswere canopy height (Anderson et al. 2007, colobus monkeys), tree basal area (Arroyo-Rodrı´guez and Mand-ujano 2006, howler monkeys), food-source abundance(Wieczkowski 2004, mangabeys), and distance to thenearest fragment (Estrada and Coates-Estrada 1996;Mandujano and Estrada 2005, howler monkeys). Thesedata were recorded for each matrix element where a troopof monkeys or solitary monkey was sighted. Canopyheight was estimated using a digital clinometer (Haglo¨f,Sweden). Food-source abundance was estimated as thenumber of trees on which leaves and fruit are knownto form part of the diet of black howler monkeys(Pozo-Montuy and Serio-Silva 2006). This number wasobtained using a 2 m  9  50 m transect in each matrixelement where monkeys were sighted (Gentry 1982). Tocalculate tree basal area, the diameter at breast heightwas measured for all trees that had at least half of theirtrunk inside the transect boundaries. Finally, we digitizedthe vegetation fragments on orthophotos from 2008(1:10,000). Using ArcView 3.3 software, we determinedthe distance to the nearest fragment as measured from theedge of the fragment to the spot where the monkeys weresighted in the matrix.Data analysisUsing the information obtained in the field, descriptiveanalyses were done on troop composition and the sex–ageclass ratio. We also did descriptive analyses of the numberof troops recorded in each type of matrix element, as wellas troop size mean and mode. We analyzed the influence of the recorded matrix attributes (canopy height, tree basalarea, food-source abundance, and distance to the nearestfragment) on monkey abundance by constructing general-ized linear models (GLMs) using the S-Plus 2000 program(Poisson log linear model) (MathSoft 1999; Crawley2002). We identified the minimum adequate modelfollowing the principles of parsimony and maximumlikelihood (Crawley 2002). Explanatory variables wereidentified in a stepwise fashion, retaining the significantvariables according to the results of the  F   tests. Beforerunning the models, each independent variable was ana-lyzed to test its degree of correlation with the other inde-pendent variables using Spearman’s rank correlation test tohelp eliminate the possibility of mistakes in the interpre-tation of the model (e.g., false positives). A value of  P  B  0.05 suggested they are correlated. Table 1  Matrix elements identified in the fragmented landscape of Balanca´n, Tabasco, MexicoName Description1. Early secondaryvegetationYoung native trees; height \ 10 m,canopy cover \ 50%2. Scrub with creepersand vinesShrubs (1–6 m) with 80–100% of theircanopy covered by creepers and vines3. Scrub with isolatedtreesShrubs (1–6 m) with isolated trees( [ 10 m)4. Pasture Pasture ( \ 1 m) with 0% tree cover5. Pasture-scrub Pasture ( \ 1 m) with scrub ( \ 3 m);horizontal cover 50–100% of thepasture’s area6. Isolated trees Native canopy tree species(height [ 10 m) separated by 3–10 m inpastures7. Living fence Native or exotic tree species(height [ 5 m) with barbed wire strungbetween them to fence in pastures8. Barbed-wire fence Fences made with barbed wire strungbetween dead wood posts to fence inpastures9. Low-growing crops  [ 0.2 ha and \ 1 m high; these tend to bewatermelon, squash, or other vine-grown species, or in some cases earlystages of crops that grow taller10. Tall crops  [ 0.2 ha and 1–3 m high; these tend to becrops that are ripe and close to harvest11. Forest plantations Exotic tree species cultivated by people(area [ 0.2 ha)12. Orchards Orchards (area [ 0.2 ha) surrounded bypastures13. Roads Roads used by people; paved orcobblestone14. Water bodies Lagoons, rivers, streams, and permanentand ephemeral wells142 Primates (2011) 52:139–147  1 3  Results Howler presence and abundance in the landscapematrixIn the landscape matrix, 244  A .  pigra  were observed, six of which were solitary males and the rest of which weredistributed among 48 groups. We also saw four that hadbeen killed by vehicles on the road. Mean troop size was5.6  ±  2.8 individuals, with a mode of three individuals.Age–sex distribution was 72 adult males, 88 adult females,32 juvenile males, 20 juvenile females, 10 male infants,and 22 female infants, including the solitary individuals.The most common group composition was multimale–multifemale with 17 troops, and unimale–multifemale with17 troops, followed by 12 unimale–unifemale troops andtwo multimale–unifemale troops. Eight of the troopsrecorded had no immatures and 16 had no infants. Theadult sex ratio was 1.3 females/male, and there were 0.3infants/adult female and 0.9 immatures/adult female.Monkey abundance by matrix element varied consider-ably. In the isolated trees, a total of 111 howler monkeys(22 groups and six solitary males) were recorded. Addi-tionally, there were 75 monkeys (11 groups) in eucalyptustrees, 27 (six groups) in orchards (  Mangifera indica,Tamarindus indica, Psidium guajava ,  Annona reticulata, Achras sapota , and  Citrus sp. ), 18 (six groups) on barbed-wire fences, and 13 (three groups) on living fences.  Alouatta pigra  activity in the landscape matrixThe monkeys were observed feeding and resting in allmatrix elements that had arboreal cover (i.e, living fences,isolated trees, orchards, and the eucalyptus plantation). Asmight be expected, on the barbed-wire fences, the onlyactivity observed was locomotion. Thirty-one percent of the troops ( n  =  15) were observed feeding. This activitywas recorded in isolated trees ( n  =  10), on living fences( n  =  1), and in the eucalyptus plantation ( Eucalyptus  spp.)( n  =  4).Matrix attributes with and without monkeysThe elements of the matrix with monkeys had the follow-ing characteristics: mean basal area of the trees in all ele-ments of the matrix was 3,487  ±  1,127 cm 2 , mean canopyheight was 13  ±  4 m, mean number of trees that form partof the howler diet (food abundance) was 6.7  ±  4, anddistance to the nearest fragment was 93  ±  87 m. The meandistance from the nearest fragment to the barbed-wirefences and roads was 72  ±  47 m. The elements of thematrix without monkeys included tall crops, low-growingcrops, pasture–scrub, pasture, scrub with isolated trees, andscrub with creepers and vines.The influence of the landscape matrix attributeson  A. pigra  abundanceA model that includes the canopy height of the matrixtrees, food-source abundance, and distance to the nearestfragment explains 74.2% of the variation in howler monkeyabundance in the matrix. According to this model, theabundance of howler monkeys increases as matrix treeheight increases and as the number of trees that form partof the howlers’ diet increases; abundance decreases withincreasing distance to the nearest fragment (Table 2).There was no correlation between the variables used in thismodel (canopy height and distance to nearest fragment: r  s  = - 0.06,  P  =  0.77; canopy height and food-sourceabundance:  r  s  = - 0.10,  P  =  0.61; food-source abundanceand distance to the nearest fragment:  r  s  = - 0.21, P  =  0.31).The model with basal area only explained 25% of thevariance in howler monkey abundance ( P  =  0.281), so thisvariable was not included in the final model (Table 2). Table 2  Parameters and fit of the minimum adequate model indicate the attributes of the landscape matrix that have a significant effect on  Alouatta pigra  abundanceAttribute  df   Slope (SE)  T P  for  t   test Changes in deviance  P  for  F   teston devianceNull model 23 276.48Intercept  - 0.403 (0.320)  - 1.25Canopy height 1 0.111 (0.019) 5.85  B 0.005 39.52 0.008Distance to the nearest fragment 1  - 0.004 (0.001)  - 4.42  B 0.005 21.14 0.044Food availability 1 0.159 (0.019) 8.22  B 0.005 144.49  B 0.005Residual deviance 20 71.32Analysis of generalized linear models (GLMs) for the attributes that affect  A. pigra  abundance. GLM/Poisson dispersion parameter  =  1(significant model  =  null deviance  =  276.48, residual deviance  =  71.32).  t   and  F   tests considered significant at [ 0.05 df   Degrees of freedom,  SE   standard error,  T   coefficientsPrimates (2011) 52:139–147 143  1 3
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x