The Use of GIS and Modelling Approaches in Squirrel Population Management and Conservation: a Review
Total Page:16
File Type:pdf, Size:1020Kb
SPECIAL SECTION: ARBOREAL SQUIRRELS The use of GIS and modelling approaches in squirrel population management and conservation: a review P. W. W. Lurz1,*, J. L. Koprowski2 and D. J. A. Wood2 1School of Biology and Psychology, IRES, Devonshire Building, University of Newcastle, Newcastle upon Tyne, NE1 7RU, UK 2Wildlife Conservation and Management, School of Natural Resources, University of Arizona, Tucson, Arizona, 85721, USA We review modelling approaches in relation to three cosmopolitan distribution. Squirrels are managed as game key areas of sciurid ecology: management, disease risk or fur-bearers that provide considerable subsistence and 5 6 assessments and conservation. Models enable us to ex- economic value , especially in Holarctic species . Tree plore different scenarios to develop effective manage- squirrels are also viewed as pests in many regions, attack- ment and conservation strategies. They may also assist ing crops, trees and electrical systems or competing with in identifying and targeting research needs for tree native species6–8. Modelling in a natural resources man- and flying squirrels. However, there is a need to refine agement context has usually focused on habitat-based techniques and assure that data used are applicable at methods and harvest dynamics. the appropriate scale. Models allow managers to make Habitat-based models have been applied to two common informed decisions to help conserve species, but suc- species of North America, eastern fox squirrels (S. niger) cess requires that the utility of the tool be evaluated as 9–11 new empirical data become available and models re- and eastern grey squirrels . Models identify habitat in fined to more accurately meet the needs of current terms of two relatively simple components: winter food conservation scenarios. availability and cover requirements. Similar approaches were applied to identify and model habitat quality of na- Keywords: Geographic information system, habitat tive Japanese squirrels (S. lis) and introduced Pallas’s suitability index, modelling approaches, population man- squirrels (Callosciurus erythraeus) from empirical data in agement, Sciuridae. Japan; the results of the modelling efforts predict marked overlap between the species12,13. Similar attempts have Introduction been made for northern flying squirrels14–16 (Glaucomys sabrinus), Siberian flying squirrels17 (Pteromys volans) MODELS have been described as simplifications of reality and Abert’s squirrels18 (S. aberti). Increasingly, these ap- that are used to help us understand complex systems1. proaches incorporate data from satellite imagery to de- Therefore, models can assist as tools in the management velop predictive models of patch occupancy, habitat use and conservation of many species. An examination of the and suitability16,19–24. literature on tree and flying squirrels (Mammalia: Sciuri- Human influence on population persistence has also dae) for application of modelling and geographic infor- been examined using a theoretical approach. Habitat suit- mation system (GIS) techniques reveals an abundance of ability index (HIS) models have been applied success- approaches (Table 1). Their uses range from statistical ap- fully to assess distribution of urban populations of eastern proaches such as generalized linear models (GLM) or grey squirrels by identifying the most important tree spe- principal component analysis (PCA) relating presence of cies and size classes, and provide guidance for urban for- tree or flying squirrels to distribution and quality of habi- esters25. Impacts of deforestation on eastern fox squirrels, tat2,3, to spatially explicit population dynamics models eastern grey squirrels, Eurasian red (S. vulgaris), and red (SEPMs) predicting eastern grey squirrel (Sciurus caro- squirrels (Tamiasciurus hudsonicus) were assessed with linensis) range expansion4. In this article, we present a regression models to identify responses to habitat frag- targeted review of key GIS and modelling studies in rela- mentation and effect thresholds26–29; Eurasian red squirrels tion to three areas of sciurid ecology: management, dis- appear particularly sensitive to fragmentation, whereas ease risk assessments and conservation. red squirrels in North America are less impacted2,29. Spa- Management tially explicit models applied to population persistence in fragmented landscapes verify that red squirrels of North Tree and flying squirrels are viewed as a natural resource America are least sensitive to fragmentation, whereas or threat to natural resources in many regions of their southern flying squirrels (Glaucomys volans) are most sensitive30. The potential for modelling approaches to as- *For correspondence. (e-mail: [email protected]) sist in ‘softening’ human impacts on tree and flying 918 CURRENT SCIENCE, VOL. 95, NO. 7, 10 OCTOBER 2008 SPECIAL SECTION: ARBOREAL SQUIRRELS Table 1. Examples of modelling approaches and applications to conservation and management of tree and flying squirrels Approach Description References Statistical models Relating habitat and other variables to presence/absence or population data 2, 14, 17, 53 using, for example, regression analysis or GLMs. Principal components analysis Summarized trends of species habitat measurements. Identifies important 3 variables in selection and reduces data dimensionality. Habitat suitability index Hypothesis of species–habitat relationships describing the landscape in 26, 27, 28 which a species could occur. Examines requirements of habitat. Spatially explicit dynamic population Individually based population dynamics models linking the spatial 4, 20, 32, 39, 42, 51, models composition of a landscape with life-history processes of mortality, 53 reproduction and dispersal. Require an understanding of the species and life-history data. General programs, e.g. RAMAS, VORTEX General population modelling tools, requiring basic life-history inputs. 46, 54 Have many extra abilities like metapopulation analysis, genetic effects, habitat analysis, etc. squirrels appears to be substantial (see ‘Conservation’ work applied the approach to Eurasian red squirrel de- below). cline in the north of England. The studies used modelling Modelling has been used to track populations due to and records of field observations to quantify the impact economic necessity or to develop viable economic strate- of the disease and to assist in the development of land- gies. Pelt production of Eurasian red squirrels is predicted scape-scale strategies for conservation43,44. by seed crops of a single tree species, two years in ad- vance5. Economic value of conservation of native Eura- Conservation sian red squirrels in the United Kingdom was examined using regression techniques to identify the amount resi- Tree squirrels are imperiled worldwide8 and several mod- dents would pay to conserve the species31. Dynamics of elling approaches have been used to aid in their conserva- introduced populations of eastern grey squirrels outside tion. Unfortunately, few life-history data exist for many their native range has been modelled to assess their im- tree squirrels45 and in such cases a general approach may plications for native fauna4,32–34. Similarly, researchers be useful. Generic averages have been used to fill life- used regression approaches to develop means of combin- history gaps and predict populations needed for the success- ing timber harvest and tree squirrel persistence17,35,36. The ful introduction of six tree squirrel species46. These use of models in evaluating management schemes has modelling efforts and empirical data reviewed therein thus often succeeded in the formulation of large-scale demonstrate that tree squirrels can establish from low management plans, suggesting the promise of using such numbers, provided introductions occur into quality habi- approaches37–39. tat and small populations can recover if factors implicated in their decline are managed46. Validation of modelling results through examination of empirical data, field ex- Disease risk assessments periments and sensitivity analyses are a necessary step to determine the utility of candidate models and to further SEPMs predict distribution of species in the landscape on modify and enhance the model value. the basis of interactions between landscape structure and For endangered Delmarva fox squirrels (S. niger ciner- life-history processes such as dispersal, reproduction and 47 20 eus), HSI models are used to identify release sites that mortality . These models are inevitably more complex have led to successful translocations48. However, recov- than statistical methods but allow for investigating species ery requires evaluation of the HSI models along with with dynamic expanding distributions, especially where mapping of habitat47. One approach to mapping habitat squirrels may act as a vector for disease. They can be used LiDAR (laser imaging detection and ranging or light used to investigate the likely spread of disease such as in detection and ranging) airborne laser to screen for appro- the case of the introduced eastern grey squirrel, which is priate vertical structure24. Sites identified as potential linked to squirrel poxvirus outbreaks in Eurasian red habitat were visited after a HSI model was used to predict 40,41 squirrels in the UK . if it matched release-site criteria. A stand-level HSI was 42 Rushton et al. integrated a SEPM with a disease also used for Siberian flying squirrels and moose (Alces model and investigated Eurasian red squirrel decline in alces).