Module 2: Wildlife Abundanceand Habitat

Margarida Santos-Reis, Reinhard A. Klenkeand Klaus Henle

Abstract The shared use of natural resources by humans and wildlife is the basisof long-lasting conflicts whose reconciliation is an urgent need. To fully under-stand the degree of damage and the conflict intensity, detailed knowledge isrequired on the distribution and abundance of animals and thus their visiting rateand contact with vulnerable resources, which in turn vary spatially depending onhabitat characteristics and regional resolution instruments. The purpose of thismodule is to assess how landscape factors and resource management factorsinfluence wildlife abundance and the exposure and vulnerability of the resource towildlife. Abundance estimates and species-habitat associations are the two keyfactors in this context. Three approaches are presented (minimum, standard andadvanced), varying in time and funding needs, and these range from literature-based educated guesses to powerful predictors using field-survey datasets, asdescribed. We suggest as a first step the standard approach and only if evidencessupport a strong influence of landscape factors select an advanced approach in alater step of the reconciliation process.

M. Santos-Reis (&)Centro de Biologia Ambiental/Departamento de Biologia Animal, Faculdade de Ciências,Universidade de Lisboa, Campo Grande, Bloco C2, 58 Piso, 1749-016 Lisbon, Portugale-mail: mmreis@fc.ul.pt

R. A. Klenke � K. HenleDepartment of Conservation Biology, UFZ—Helmholtz Centre for Environmental Research,Permoserstraße 15, 04318 Leipzig, Germanye-mail: reinhard.klenke@ufz.de

K. Henlee-mail: klaus.henle@ufz.de

R. A. Klenke et al. (eds.), Human-Wildlife Conflicts in Europe,Environmental Science and Engineering, DOI: 10.1007/978-3-540-34789-7_12,� Springer-Verlag Berlin Heidelberg 2013

231

1 Rationale and Objectives

Landscape factors influence the distribution and abundance of animals and thus theirvisiting rate and contact with vulnerable resources. Therefore, conflicts may stronglydiffer from landscape to landscape and from location to location within a landscape. Thepurpose of this module is to assess how landscape factors and resource managementfactors influence wildlife abundance and the exposure and vulnerability of the resourceto wildlife. Together with modules 3 (Damage assessment), 5 (Economic and policyanalysis), and 6 (Stakeholder analysis) it will provide spatially differentiated informationabout the degree of damage and the conflict intensity.

The module ‘‘Wildlife abundance and habitat’’ is important whenever thescreening process indicates strong spatial differentiation of the conflict and locallyor regionally differentiated resolution instruments are available or a feasibleoption. It helps addressing the following questions:

• How abundant is the wildlife species and how is their distribution within theconflict area?

• How frequent is its contact with the resource?• Are landscape factors influencing the local status of wildlife and, therefore,

influencing the conflict, e.g., by promoting or preventing losses of commerciallyimportant fish or damage to fishing gear?

• Do we have local problems triggered by conditions and processes at another siteor scale?

Different landscapes are not equally attractive and do not have the same car-rying capacity, e.g., rivers, streams, or small and large ponds in the case of fishpredators. Characteristics of water bodies (e.g., water surface of ponds, percentageof bank cover, fish density) and of the surrounding area (e.g., land cover, density ofwater bodies) explain spatial differences in wildlife occupancy with implicationsfor the intensity of the conflict. In the marine area the concept of landscape factorsis not yet as well developed as for terrestrial habitats, but like in these, structure ofthe sea floor, currents, ice cover, distance to breeding sites, etc. are correspondingfactors that may influence the distribution of the wildlife and/or the resource andcause spatial differences in conflicts. The same applies to estuarine areas whereadditional factors, such as the salinity gradient and the distance to freshwatersources, are spatially differentiated and may influence both the fish-eaters and thewild fish species that may act as important alternative prey to the former.

2 Methods and Approaches

Abundance estimates and species-habitat associations are the two key factors thatmust be assessed in order to understand how landscape factors and resourcemanagement factors influence the exposure and vulnerability of the resource towildlife.

232 M. Santos-Reis et al.

Tab

le1

Lis

tof

pote

ntia

lm

etho

dsto

beap

plie

dfo

ra

min

imum

,st

anda

rd,

and

adva

nced

appr

oach

Par

amet

ers

App

roac

hes

Min

imum

Sta

ndar

dA

dvan

ced

Dis

trib

utio

nL

iter

atur

eW

ildl

ife

surv

eyW

ildl

ife

surv

eyA

bund

ance

Exp

ert

gues

sP

rese

nce/

abse

nce

data

orre

lati

veab

unda

nce

Pop

ulat

ion

size

Vis

itin

gra

teT

radi

tion

alkn

owle

dge

Sta

ndar

dfi

eld

surv

eys

(dir

ect

obse

rvat

ions

orse

arch

ofsi

gns

ofpr

esen

ce)

Adv

ance

dfi

eld

surv

eys

(dir

ect

obse

rvat

ions

orvi

deo

surv

eill

ance

)H

abit

atav

aila

bili

tyL

iter

atur

eQ

uant

ifica

tion

from

map

(e.g

.,C

OR

INE

land

cove

rm

ap)

Rem

ote

sens

ing

tech

niqu

es(e

.g.,

aeri

alsu

rvey

s,sa

tell

ite

imag

ery)

,m

appi

ngof

spec

ial

habi

tat

stru

ctur

esK

eyla

ndsc

ape

fact

ors

Exp

ert

gues

sA

naly

sis

ofla

ndsc

ape

com

posi

tion

and

use

ofde

scri

ptiv

est

atis

tica

lm

etho

dsfo

ran

alys

isof

map

san

dsi

tedi

stri

buti

ons

(hab

itat

mod

els

base

don

fiel

dsu

rvey

s—H

abit

atS

uita

bili

tyIn

dexe

s)

Ana

lysi

sof

land

scap

eco

mpo

siti

onan

dco

nfigu

rati

onas

wel

las

spat

ial

dist

ribu

tion

and

neig

hbou

rhoo

dre

lati

onof

map

san

dsi

tedi

stri

buti

ons;

anal

ysis

ofpo

pula

tion

deve

lopm

enti

nti

me

and

spac

e;us

eof

infe

rent

ial

stat

isti

csfo

rhy

poth

esis

test

ing

(hab

itat

mod

els

base

don

capt

ure/

reca

ptur

e,te

lem

etry

—G

ISba

sed

mod

els)

Module 2: Wildlife Abundance and Habitat 233

2.1 Minimum Approach

A minimum requirement should be a review of readily accessible literature ondistribution, habitat requirements, and habitat preferences of the conflict species(see Table 1). In addition, expert knowledge may be consulted with the advantagethat experts are usually familiar with the relevant literature, but having the dis-advantage that they may belong to specific interest groups within the conflict. Ifavailable data prove to be insufficient or inadequate to assess key factors, field-work, as outlined in the standard approach, is mandatory.

2.2 Standard Approach

A standard approach, which will suffice for most purposes, will be to collect dataon the presence and relative abundance of the wildlife species and on landscapeand human resource factors (e.g., fish stocking densities) assumed to be importantfor the attraction of the wildlife to the resource (see Table 1). The extent ofdamages depends on a mixed set of factors (e.g., availability of alternative prey,multi-species-interactions), more than just the distribution and abundance of thewildlife species. However, the contact with vulnerable resources (visiting rate)

Fig. 1 Well trained professional or voluntary fieldworkers are mandatory for estimation andmonitoring of abundance even with simple methods using sight counts or signs of presence.Photo: André Künzelmann

234 M. Santos-Reis et al.

may contribute to the understanding of the species occupancy pattern and mayexplain perceived damage.

Measurements done by direct observation have the advantage of providing thenumber of animals visiting the resource, but may be time consuming or impossible(e.g., for carnivores), and the mere presence of the observer may influence thevisiting rate (Fig. 1). The number of signs or frequency with which they areobserved in repeated surveys, provides a relative index of visiting rates, notabsolute numbers of animals visiting the resource. In most cases this should besufficient. When using signs of presence, it is mandatory to have trained field-workers. For methodological recommendations to assess visiting rates see Box 1.

Box 1 Visiting Rates and Population Size

Visiting rates are a measurement of the number of times a species encountersthe resource during a certain time span (e.g., day, week). This may be donethrough direct observation of the animals (e.g., birds) or indirectly bysearching specific signs of presence (e.g., carnivores) or by assessingdamage caused to fishing gear (e.g., marine mammals). Measurementsshould be repeated systematically in order to have a reliable value. It mightalso be important to replicate this measurement in time (e.g., seasonally) ifthe biology of the species indicates that its presence varies significantly withtime (e.g., migratory species, species with changes in habitat use duringbreeding). Replicates will allow a better assessment of wildlife impacts oncontested resources.The search of tracks and/or faeces is a common field method when studyingcryptic animals, such as the otter. In the frame of the FRAP project, ottervisiting rates to selected fish ponds were assessed based on repetitive moni-toring of old and fresh spraints. Ponds were checked weekly and the presence(or absence) of fresh (from previous night) and old (from the other six nights)spraints was recorded. A new maximum likelihood approach was developed toestimate the probability of the species visiting the resource per day. Resultsfrom a simulation experiment show that the estimator provides an unbiasedand precise estimate if the sampling effort is strong (5–20 weeks) and theweekly visiting rates lie between 0.1 and 0.6.In a more advanced approach, field surveys may assess visiting rates withvideo surveillance and supply the number and behavior of animals involvedin those visits. Important limitations are that video-surveillance is expensive,time-consuming in the tape analysis, and cannot be applied to a large scale,besides needing to be optimized case by case (e.g., number and position ofvideo cameras). It may be, however, the only solution to understand thereason why in specific situations high visiting rates, assessed through scatand footprint counting, do not correlate with the damage extent in terms ofdiet (e.g., otters in Sado Estuary, Portugal—Sales-Luís et al. 2009) and viceversa.

Module 2: Wildlife Abundance and Habitat 235

Visiting rates may be used as a surrogate of population abundance but do notnecessarily correlate to population size, a parameter required in advancedapproaches. For some species population size can be calculated preciselythrough direct observation (e.g., birds at breeding colonies) and regularmonitoring provides time series data on trends, most useful in advancedapproaches. For others, however, to obtain even a relative index of abun-dance (standard approach) is a rather difficult task, as is the case of mostcarnivores that are mainly solitary, nocturnal, extremely elusive, and live atlow densities. Absolute numbers, necessary for advanced approaches, can beestimated only under special environmental conditions (e.g., new snowcover) or taking profit of recent advances in molecular tools using scats(molecular scatology). Molecular scat analysis is a costly and time-con-suming technique that still has a relatively low rate of success in samplescollected in adverse environmental conditions (e.g., extremely high tem-peratures or level of humidity) and/or incorrectly preserved until analysis,two factors that contribute to DNA degradation. As a methodologicalguideline for estimating relative and absolute numbers of wildlife species,Caughley (1980) is still the best reference for non-specialist users. For userswith more advanced statistical skills, Williams et al. (2002) is recommended.For species specific survey methods Cooperrider et al. (1986) is a very usefulresource. The estimation of visiting rates, as developed and used in theFRAP-project, is described in detail by Gruber et al. (2008).

Based on distribution pattern and habitat requirements, simple species-habitatassociation models can be developed. The advantage of habitat modelling pro-cedures is that the probability of presence or relative abundance of the wildlifespecies can be extrapolated to larger areas and distribution maps can be producedquickly and efficiently, especially if inexpensive environmental data that coverlarge regions are available. See Box 2 ‘‘Habitat analysis and habitat modeling’’ fortechnical and statistical recommendations. Limitations do exist, however, andthese are related to the scale and quality of the datasets (e.g., whereas data may bevery concrete and easily available for some resources, such as fish in small win-tering ponds, they may be extremely difficult to collect in some cases, such as inmarine fish resources) as well as to the time and funding available. Also, whenextrapolating across large spatial scales, one needs to keep in mind that theimportance of landscape factors may change in different types of landscapes.

Box 2 Habitat Analysis and Habitat Modeling

Presence/absence data, either geo-referenced or mapped over a grid system,can be associated to a series of environmental descriptors. They are widelyused to develop habitat models that allow extrapolation to non-sampledareas. Various statistical methods exist to derive habitat models from fielddata, such as simple habitat suitability models, which group environmental

236 M. Santos-Reis et al.

data into classes based on species distribution data, or multivariate statisticalmodels (logistic regression, discriminant analysis, Generalized LinearModels—GLM, Generalized Additive Models—GAM) to predict the loca-tion and extent of habitat for wildlife species. Recent developments in thefield of computational ecology allow the use of mixed ecological datasets,including individual capture/recapture and radio-tracking data. GeographicInformation Systems (GIS) facilitate an optimal use of existing spatial data,e.g., on land cover, soils, geology, or topography. By combining mathe-matical and computer models of natural systems with geographically explicitdetails of the biotic and abiotic components of the environment, theunderstanding of natural systems and the ability to properly manage it and topredict the effects of alternative management options is improved.Most methods require species presence and absence data, with presencesattesting a suitable habitat and absences an unsuitable habitat. An overviewof relevant statistical methods can be found in Corsi et al. (2000). There is,however, an important constraint, which is the fact that blank areas areambiguous as to whether the species is really absent or no records wereavailable. Recent developments in this area include modeling species habitatsuitability with presence only data, a new area that deserves to be furtherexplored (e.g., Hirzel et al. 2002; Pearce and Boyce 2006). Model validationtechniques are however needed to ensure that model predictions are notstatistic artifacts.The sampling points selected for the development of habitat models in thefirst step should cover the whole range of variation of the landscape factorsand should be spaced as far apart as feasible to avoid autocorrelation andpseudo-replication.The advantage of habitat modeling procedures is that, if inexpensive envi-ronmental data covering larger regions are available, the probability ofpresence or relative abundance of the wildlife species can be estimated anddistribution maps can be produced quickly and efficiently.

2.3 Advanced Approach

If the screening process indicates that the conflict may have very strong spatiallydifferentiated components and spatially differentiated mitigation or managementinstruments may be feasible, one should consider the advanced approach, whichconsists of estimating the population size (density) of the wildlife species (Box 1),which then can be related to landscape factors using more sophisticated statisticalmethods (Box 2). The calculation of real loss, based on consumption rates of theresource by wildlife, also requires knowledge of absolute numbers of wildlifevisits to the resource (modules 3 and 5) obtainable only with advanced approaches.Finally, knowledge of the population size of the wildlife species or the carrying

Module 2: Wildlife Abundance and Habitat 237

capacity of the habitat is usually needed for the modeling of the viability of thewildlife species under different management scenarios. Such more in-depthanalyses are usually called for if the adequacy of financial compensations for lossof resources to wildlife may be a key driver of the conflict.

3 Recommendations and Conclusions

Each wildlife species is associated with key landscape factors that are undoubtedlydriving its distribution and density at a certain geographic area. These associationsmay change through time (e.g., seasonally) or space (within its full range) and inturn have a differentiated impact upon the extent of damage the wildlife speciesmay cause. Responses of wildlife to changing landscapes or human resource usefurthermore influence population viability (module 8) and determine which tech-nical mitigation measures are adequate, and where and when they have to be used(module 7). The use of spatially differentiated approaches and the coordination ofthese with mitigation measures (module 7) and policy instruments (module 9) arecrucial to mitigate conflicts (Reimoser 2002).

Therefore, it is mandatory to have updated information on the distribution andstatus of the wildlife species involved in the conflict. Table 1 summarizes potentialapproaches. This may be comparably easy with terrestrial species that are easilyseen and identified (e.g., cormorants) or, although elusive, do leave veryconspicuous and characteristic signs of presence (e.g., otters), but may be moredifficult with wide-ranging marine species (e.g., seals) that need appropriatetechnical means, leaving as an option the carrying out of oral inquiries to relevantinterest groups. However, the presence of a predator at a given conflict area alonedoes not provide sufficient information about the extent of damage the species maycause. To accurately predict the predatory impact, an abundance estimator ishighly recommended. Varying with the species body size, life-style (solitaryvs. gregarious), activity pattern (diurnal vs. nocturnal), and habitat requirements(generalist vs. specialist), this is usually difficult, time-consuming, and needshighly qualified personnel, being however the only option in some situations.

Complementarily, to identify key landscape features that are influencing theoccupancy pattern and abundance of the wildlife species, the use of the freelyavailable CORINE land cover map is recommended as the best possible approachwith limited resources. More expensive (high resolution maps) or time-consuming(field surveys) methods should be restricted to situations where conflict recon-ciliation clearly requires a spatially differentiated approach.

Identification of landscape factors promoting damage to fish farms requiresgood datasets. As outlined in module 3 (Damage assessment), wildlife may have alarge local impact, either directly (fish eaten, damaged or disturbed) or indirectly(damage to catch and gear) (Freitas et al. 2007; Santos-Reis et al. 2013). Land-scape factors that contribute to the carrying capacity of an area may be useful toassess the risk of impact to fisheries. These, however, act at different scales

238 M. Santos-Reis et al.

according to the species and the area. For migratory species (e.g., birds) and forwide ranging marine species (e.g., seals) the landscape scale seems to be theappropriate one, although the relevant factors may be difficult to manage directlyand indirect measures may be needed. For small sized terrestrial species, however,local scale factors better explain damages, and relevant factors may be modified toreduce predation.

In summary, the minimum approach only provides an educated guess about thekey landscape factors that may be influencing wildlife numbers and distribution,and hence about the potential extent of damage (low, medium, or high). With thestandard approach the reliability of the results provides a much better insight intothe conflict. Advanced approaches result in more powerful predictors but are moreexpensive, more time-consuming, and require specialized manpower. Specificallytargeted research projects may be essential for an adjustment of methods to thespecific problem at hand. The obvious way to handle these trade-offs is to use as afirst step the standard approach and then, if the results reveal that landscape factorsare crucial in conflict reconciliation, select an advanced approach in a later step ofthe reconciliation process or try to get the support of a research organization thathas the expertise and scientific interest to pursue an advanced approach.

References

Caughley G (1980) Analysis of vertebrate populations. Wiley, New YorkCooperrider AY, Boyd RJ, Stuart HR (1986) Inventory and monitoring of wildlife habitat. United

States Department of the Interior, Bureau of Land Management, Service Centre, DenverCorsi F, de Leeuw J, Skidmore A (2000) Modelling species distribution with GIS. In: Boitani L,

Fuller T (eds) Research techniques in animal ecology: controversies and consequences.Columbia University Press, New York, pp 389–434

Freitas D, Gomes J, Sales Luís T, Madruga L, Marques C, Baptista G, Rosalino LM, Antunes P,Santos R, Santos-Reis M (2007) Otters and fish farms in the Sado Estuary: ecological andsocio-economic basis of a conflict. Hydrobiologia 587:51–62

Gruber B, Reineking B, Calabrese JM, Kranz A, Poledníková K, Poledník L, Klenke R, ValentinA, Henle K (2008) A new method for estimating visitation rates of cryptic animals viarepeated surveys of indirect signs. J Appl Ecol 45:728–735

Hirzel AH, Hausser J, Chessel D, Perrin N (2002) Ecological-niche factor analysis. How tocompute habitat-suitability maps without absence data. Ecology 83:2027–2036

Pearce JL, Boyce MS (2006) Modelling distribution and abundance with presence-only data.J Appl Ecol 43:405–418

Reimoser F (2002) Wildlife ecological spatial planning (WESP): an instrument for integratingwildlife into comprehensive land management. In: Proceedings of the XXIVth congress of theinternational union of game biologists (Thessaloniki, 1999), pp 176–185

Santos-Reis M, Santos R, Antunes P, Sales-Luís T, Gomes J, Freitas D, Madruga L (2013)Reconciliation of the conflict between otters and fish farmers: lessons from Sado estuary inPortugal. In: Klenke R, Ring I, Kranz A, Jepsen N, Rauschmayer R, Henle K (eds) Human-wildlife conflicts in Europe—fisheries and fish-eating vertebrates as a model case. Springer,Heidelberg, doi:10.1007/978-3-540-34789-7_4

Santos-Reis M, Santos R, Antunes P, Sales-Luís T, Gomes J, Freitas D, Madruga L (2013)Reconciliation of the conflict between otters and fish farmers: lessons from Sado estuary in

Module 2: Wildlife Abundance and Habitat 239

Portugal. In: Klenke R, Ring I, Kranz A, Jepsen N, Rauschmayer R, Henle K (eds) Human-wildlife conflicts in Europe—fisheries and fish-eating vertebrates as a model case. Springer,Heidelberg, pp 30–30

Williams BK, Nichols JD, Conroy MJ (2002) Analysis and management of animal populations.Academic Press, San Diego

Free and Commercial Software for Statistical Analysis and GIS

ArcGIS—ESR. http://www.esri.com. Accessed 08 July 2009BioMapper. http://www2.unil.ch/biomapper/. Accessed 08 July 2009GRASS GIS—The world leading free software GIS. http://grass.fbk.eu. Accessed 08 July 2009R Development Core Team (2006) R: a language and environment for statistical computing. R

foundation for statistical computing, Vienna, Austria. http://www.r-project.org. Accessed 28Aug 2006

SAGA GIS—A system for an automated geographical analysis. http://www.saga-gis.org/en/index.html. Accessed 27 June 2012

Web Pages

CRAN Task View: Analysis of ecological and environmental data. http://cran.r-project.org/web/views/Environmetrics.html. Accessed 08 July 2009

European Environmental Agency (EEA) dataservice. http://dataservice.eea.eu.int/dataservice.Accessed 08 July 2009

Free GIS software and data. http://www.freegis.org. Accessed 08 July 2009Habitat Evaluation Procedures Handbook. http://www.fws.gov/policy/ESMindex.html. Accessed

08 July 2009Patuxent Software Archive. http://www.mbr-pwrc.usgs.gov/software.html. Accessed 08 July

2009phidot.org home page. http://www.phidot.org. Accessed 08 July 2009USGS National Wetlands Research Center: Habitat Suitability Index Models Introduction. http://

www.nwrc.usgs.gov/wdb/pub/hsi/hsiintro.html. Accessed 08 July 2009

240 M. Santos-Reis et al.