当前位置：文档库 › comparing the effects of bird diversity conservation approaches on landscape

comparing the effects of bird diversity conservation approaches on landscape

A hybrid scheme for comparing the effects of bird diversity conservation

approaches on landscape patterns and biodiversity in the Shangan sub-watershed in Taiwan

Chen-Fa Wu a ,Yu-Pin Lin b ,Shin-Hwei Lin c ,*

Department of Horticulture,National Chung-Hsing University,No.250,Kuo Kuang Rd.,Taichung 402,Taiwan

Department of Bioenvironmental Systems Engineering,National Taiwan University,No.1Sec.4Roosevelt Rd.,Taipei 106,Taiwan c

Department of Soil and Water Conservation,National Chung-Hsing University,No.250Kuo Kuang Rd.,Taichung 402,Taiwan

a r t i c l e i n f o

Article history:

Received 12May 2010Received in revised form 9December 2010

Accepted 1March 2011

Available online 29March 2011Keywords:

Watershed planning Landscape ecology Bird diversity

Land-use change modeling Landscape metrics

a b s t r a c t

This work utilizes bird survey data,regression modeling,land-use modeling and landscape metrics to evaluate the effects of various spatial bird diversity conservation approaches on land-use allocation,land-use patterns,and biodiversity in the Shangan sub-watershed in central Taiwan.A survey of the distribution of species revealed that bird species are concentrated in the central and western parts of the sub-watershed.The results obtained using a Shannon e Weaver diversity regression model suggest that diversity of land-use increases the diversity of bird species.Logistic regression results verify that socio-economic factors determine the potential advantages of designating a particular type of land-use in certain parts of the study area.The results of land-use simulation modeling indicate that the eastern and southwestern areas of the sub-watershed will change most frequently between 2007and 2017.Addi-tionally,increasing the areas to protect bird diversity will effectively increase the patch size,habitat core area,edge effect and habitat connectivity.The Shannon e Weaver diversity regression model shows that protecting bird species diversity in large areas increases bird diversity.The proposed modeling approach is an effective tool that provides useful information for ecological planning and policymaking related to watersheds.

1.Introduction

Birds are excellent indicators of environmental change because they are easy to monitor,high in the food chain and very sensitive to changes in the structure and composition of their habitat (Devictor and Jiguet,2007;Lin et al.,2008a ).As a result,they are often used in the ecological monitoring and assessment of urban (Fernández-Juricic,2004;Lin et al.,2008b ),suburban (Woodhouse et al.,2005),rural (Pino et al.,2000;Yamaura et al.,2005;Vesk and Nally,2006)and reserve (Chapman and Reich,2007)areas.Inwatersheds,changes in land-use have been the primary cause of declining bird diversity in recent decades (Tucker et al.,1994;Fernández-Juricic,2004).Current land management protocols emphasize the sustainable diversity of landscape scales,and increasingly recognize the ecological value of watersheds.Indeed,such watersheds not only link natural areas,but also contribute to several ecological processes (Felton,1996;Pino et al.,2000).Therefore,land-use planning and management can

increase the ecological value of a watershed,and thus constitute a major ?eld of research.

Simulations have been used to elucidate biodiversity issues asso-ciated with watershed management,including the effect of conser-vation policies on bird habitats and landscape patterns (M?rtberga et al.,2007).An integrated landscape model has the potential to predict the effects of management practices and land-use patterns on the environment (Turner et al.,2001).Therefore,the development of an integrated method for simulating and assessing land-use changes,land-use patterns and their effects on bird habitats and landscape patterns at the watershed level is critical to biodiversity conservation,land-use planning and watershed management.Various modeling approaches have been used to simulate the patterns and conse-quences of land-use changes under various land-use management policies.The diverse models used to elucidate land-use changes include stochastic models,optimization models,dynamic process-based simulation models and empirical models (Verburg et al.,2002;Lin et al.,2007,2009;Castella et al.,2007).One land-use model that has been employed extensively is the Conversion of Land use and its Effects model (CLUE-s),which was developed to simulate land-use changes by considering empirically quanti ?ed relationships between

*Corresponding author.Tel./fax:t886422854322.E-mail address:shlin@https://www.wendangku.net/doc/fa7648139.html,.tw (S.-H.

Lin).

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage:www.elsev https://www.wendangku.net/doc/fa7648139.html,/locat

e/jenvman

Journal of Environmental Management 92(2011)1809e 1820

land-use and its causal factors in combination with dynamic modeling(Verburg et al.,2002;Verburg and Veldkamp,2004).A number of recent studies have employed CLUE-s to simulate land-use changes in watersheds(Lin et al.,2007,2008a,2008b,2009).

Landscape ecology provides a conceptual framework for devel-oping ecological networks because it emphasizes spatial and temporal dynamics and yields a larger-scale view than provided by traditional site-based conservation(Forman and Godron,1986; Forman,1995;White et al.,1997;Pino et al.,2000).In a water-shed,the suitability of a habitat for birds is affected by the land-scape pattern and type of land-use.Changes in land-use or landscape patterns may affect some species positively and others negatively.Hence,to evaluate the suitability of a habitat, researchers must know how bird species respond to habitat factors on coarse spatial scales and to patterns of land-use within a certain radius of their habitats(Graf et al.,2005;Holzk?mper et al.,2006). Numerous works have developed models to simulate the suitability of a particular habitat for birds,e.g.,the stepwise multiple regres-sion model(Rafe et al.,1985;Pino et al.,2000)and the logistic regression model(Holzk?mper et al.,2006;Scozzafava and De Sanctis,2006;M?rtberga et al.,2007).The stepwise multiple regression model is useful for evaluating the relationships between the diversity or richness of species and environmental variables, whereas the logistic regression model is normally used to examine the relationships between the presence/absence of speci?c species and environmental variables.

The landscape structure usually interacts with various ecological processes(Forman and Godron,1986;Forman,1995),so spatial analysis of a landscape may be effective in elucidating underlying ecological relationships(Turner and Gardner,1993;Miller et al.,1997). Landscape metrics are frequently used to evaluate changes in land-scape patterns and the characteristics of bio-habitats.The composi-tion,con?guration and connectivity of a landscape are primary descriptors of landscape patterns and bio-habitat characteristics (Turner et al.,2001).Landscape patterns and bio-habitat character-istics can be quanti?ed using spatial landscape indices or metrics that quantify landscape composition,con?guration and connectivity.The spatial con?guration of a landscape refers to the spatial character and arrangement,position,or orientation of patches in a class or land-scape.The composition of a landscape refers to the features that are associated with the diversity and abundance of patch types within a landscape.The connectivity of a landscape refers to the features that are associated with the distance between a patch and its nearest neighbor in a landscape(McGarigal and Marks,1995).Metrics that represent landscape patterns,which are themselves speci?ed by composition,con?guration and connectivity,include the number of patches,area,shape,core area,nearest-neighbor distance,landscape diversity,interspersion and contagion metrics.A number of recent studies have integrated landscape structure and heterogeneity vari-ables with land cover and distribution patterns to evaluate bird species diversity(e.g.Pino et al.,2000;Luoto et al.,2004;Bennett et al.,2004;Oja et al.,2005;Langanke et al.,2005).

This work proposes a hybrid approach for comparing the effects of bird diversity conservation schemes on landscapes and on the biodiversity of a sub-watershed.The hybrid method combines a bird survey with regression modeling,modeling of land-use patterns and analysis of the land-use patterns(landscape metrics).

A regression model of the suitability of a habitat for a diverse range of birds was established using data obtained in a survey of bird, land-use derived from1:5000aerial photographs taken in2007. The results of the regression analysis were used to construct?ve scenarios of bird habitat conservation.CLUE-s,the simulation model of changes in land-use,was then applied to simulate changes in land-use from2007to2017under the?ve https://www.wendangku.net/doc/fa7648139.html,ndscape metrics were applied to evaluate the effect of conservation policies on landscape patterns.Finally,the total value of the predicted bird diversity probability was calculated to evaluate the ecological effects of bird diversity under the?ve scenarios.

2.Material and methods

2.1.Study area

The Shangan sub-watershed in the Chenyulan river watershed in central Taiwan has a total area of209ha(Fig.1).It is a transition area between a high mountainous area in the east and a?oodplain in the https://www.wendangku.net/doc/fa7648139.html,nd cover is dominated by forest and land prone to landslides in the east,agricultural land in the middle,and water body,betel nut cultivation and agriculture in the west.Human settlements are concentrated in non-mountainous areas on the ?oodplain.The Shangan sub-watershed is a typical mountainous-type drainage watershed that is elongated in a west e east direction with an average altitude of729.2m and a slope of20.9 .More than 73%of its tributaries have gradients greater than30%.

The epicenter of the Chi e Chi earthquake that occurred in southern Nantou County in1999is close to the Shangan sub-watershed.The earthquake caused surface ruptures approximately 100km long in the north e south direction of the Chelungpu fault and 10,000landslides,which caused signi?cant changes in landscape patterns throughout central Taiwan,especially the area close to the epicenter(Lin et al.,2006).The following year,typhoon Xiangsane struck with a maximum wind speed of138.9km/h,a radius of 250km,and gusts of166.7km/h.The maximum daily rainfall was 550mm/day.Then,on July30,2001,typhoon Toraji struck with a maximum wind speed of138.9km/h and a180km radius.The typhoon brought extremely heavy precipitation(230e650mm/day) and triggered more than6000landslides in Taiwan.Natural distur-bances in the form of earthquakes and typhoons generally in?uence the isolation,size,and shape complexity of patches in the landscape and the class levels in watersheds in Taiwan(Lin et al.,2006).

The earthquake and typhoons caused major changes in landscape patterns in the Shangan sub-watershed and especially in forest,grass, farmland and bare land.Steep slopes,loose rocks and continuous heavy rainfall were the apparent causes of the heavy slump sand debris?ows.Improper land-use is a major determinant of the magnitude and frequency of natural hazards like landslides.After such disturbances,landscape restoration and watershed manage-ment become priority tasks for Taiwan’s Soil and Water Conservation Bureau(S.W.C.B.).From2001to2006,the S.W.C.B.spent more than NT$382million on restorative landscape projects in the Shangan sub-watershed(Soil and Water Conservation Bureau,2007).Maximizing the ecological effectiveness of these efforts has become an important issue in the management of the sub-watershed.

2.2.Collection of data on birds and species diversity

Point counting is one of the most widely used methods for surveying bird species(Buckland et al.,2001).Cost of time and money often in?uences times and periods of bird survey.So,many researches had been used short collectionperiod for bird data to evaluate the bird diversity(Vidaurre et al.,2006;Smith and Wachob,2006;Goetz et al., 2007)and predict long-term bird species distribution(Mitchell et al., 2008).For this study,bird species in the Shangan sub-watershed were surveyed by?eld sampling in autumn and winter2006and spring and summer of2007(Fig.1).The study area was divided into 1km?1km squares using the Universal Transverse Mercator system. Fifty samples were taken if accessibility and geographical conditions permitted.In each sampling,one transect was made for10min in each season by a researcher who walked100m around a survey area (Scozzafava and De Sanctis,2006).During these transects,visual and

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e1820 1810

Fig.1.2007survey of bird species,land use,and sub-watershed.

C.-F.Wu et al./Journal of Environmental Management 92(2011)1809e 18201811

audio data about numerous bird species was collected.In the study,36%of sampled points were not along the roads or built-up area.Moreover,in each sampling that one transect was made for 10min by a researcher who walked 100m around a survey area.Fifty survey areas were not overlay which is useful to reduce bias of result of the analysis (Fig.1).

To compare areas with high diversity of bird species,the Shannon and Weaver diversity index was calculated for each season to generate the index (Eq.(1))of species diversity for each sample (Magurran,1988).The mean Shannon and Weaver diversity index across the four seasons was then used as a dependent vari-able in stepwise multiple regression,while the environmental variables and landscape index were treated as independent variables.

H *?à

p i ln p i

(1)

where p i is estimated as n i /N ;n i is the number of the i th species as

a proportion of the total population,and N ?P

n i .2.3.Model of bird species diversity

Predictive modeling,rather than using scattered plots to study the number of birds requires large study areas so that variations can be taken into account (Luoto et al.,2004).Stepwise multiple regression analysis was adopted to evaluate the signi ?cant effects of land-use,the proportion of an area associated with a particular type of land-use,physical pattern variables,and landscape pattern variables on the spatial diversity of bird species.Stepwise multiple regression analysis has been performed in similar studies (e.g.,Rafe et al.,1985;Pino et al.,2000;Luoto et al.,2004)to elucidate the contribution of independent variables to the total variance in the regression.In the proposed regression model,the mean value of the four-season Shannon and Weaver diversity index is a depen-dent variable.Factors that in ?uence the spatial diversity of bird species include proportional land-use parameters,physical factors,and landscape pattern factors within a 100m radius scope.

To reduce the number of independent variables and thereby increase the robustness of the analysis,landscape units are cate-gorized as forest,betel nut cultivation,agriculture,grassland,built-up areas,water bodies and landslide areas (Pino et al.,2000).When apply 1:5000aerial photographs to digitalize land-use map that accuracy should be over 95%(National Land Surveying and Mapping Center,2006).In this study,land-use data was obtained from a digital land-use map that was based on 1:5000aerial photographs taken in 2007.For the digital accuracy consider that follow researchers take 1:5000aerial photographs to make sure land-use type at each patch in the study area.The total classi ?ca-tion accuracy reaches 97.6%.

Physical factors include altitude and slope,as well as distance to roads,rivers and built areas.The DTM,which has a spatial resolution of 40m,was applied to calculate the altitude and slope of the study area.Spatial analysis was performed using Arcview 3.2to calculate the distances from the land-use unit to the road,river and the nearest built-up area,using the above-mentioned land-use map.

To determine how landscape patterns in ?uence bird diversity in the study sub-watershed at the landscape scale,landscape metrics along a 100m perimeter of the sample site were calculated by patch analysis (McGarigal and Marks,1995)using the GIS software package,Arcview 3.2.The following six landscape indices were considered:the number of patches (NP),the mean patch size (MPS),the total number of edges (TE),the edge density (ED),the mean patch edge (MPE),and the mean shape index (MSI).The provided FRAGSTATS (McGarigal and Marks,1995)user menu describes the landscape indices and their equations in detail.

2.4.Model of land-use change

The two parts of the Conversion of Land use and its Effects (CLUE-s)model are a non-spatial demand module and a spatially explicit allocation procedure.The non-spatial module calculates the change in the total area of each land-use type in the aggregate (Verburg et al.,2002).In the spatial explicit allocation procedure,non-spatial demands are transformed into land-use changes at various locations in the study area.Yearly land-use demands,which must be speci ?ed before the allocation,can be set by various approaches.The allocation is determined by the results of empirical and spatial analyses as well as dynamic modeling (Verburg et al.,2002).

The empirical analysis of a location ’s suitability ?rst requires the collection of relevant data.In addition to data on land use,data that speci ?es the other factors that drive changes in land-use were also collected (Lambin et al.,2001).The relationships between land-use types and the respective driving factors were evaluated by stepwise logistic regression using the following equation (Verburg et al.,2002):

Log

P i

1àP i

?b 0tb 1X 1;i tb 2X 2;i t/tb n X n ;i ;(2)

where P i is the probability that a grid cell is associated with the land-use type of interest;X is a driving factor;and b i is the coef ?cient of a driving factor in the logistic model.The goodness of ?t of the regression models was evaluated using the Relative Operating Characteristic,which is de ?ned as the area under the curve that relates the proportion of true positives to the proportion of false positives for an in ?nite number of cut-off values (Overmars and Verburg,2005).Hence,the values of the Relative Operating Char-acteristic vary between 0.5(completely random)and 1(perfect discrimination).In this study,forward stepwise logistic regression and Relative Operating Characteristic analyses were performed using the Statistical Package for the Social Sciences (SPSS)for Windows (SPSS Inc.,Illinois,USA).Probability maps of all land-use types were produced from the logistic regression results.

Next,the spatial policies,conservation area,and decision rules (including a land-use transition matrix)were speci ?ed for the studied watershed.For each land-use type,a specify conversion elasticity was calculated to determine the typical conversion conditions of different types of land-use (Verburg et al.,2002).The model employed an iterative procedure to allocate changes in land-use based on probability maps,decision rules combined with actual land-use maps,and the demand for the various types of land-use (Verburg et al.,2002).The digital land-use map was based on 1:5000aerial photography obtained in 2007,and land-use classi-?cations included forest,built-up land,agricultural land,grassland,water,betel nut land and landslide land.The factors that drove land-use changes in the study area included land ownership,soil erosion,soil depth,the proportion of organic matter in the soil,soil pH,the soil erosion,altitude,slope,distance to rivers,distance to roads,and distance to soil and water conservation area.Data sources and resolution information of land-use and driving factors were collected in Table 1.All of these factors and land-use data were converted into a grid with the same resolution of 40m.

The data for all land-use types and the driving factors was input to the logistic regression model to formulate land-use change probability maps for each land-use type.The modeling procedure is described in detail in Verburg et al.(2002)and Verburg and Veldkamp (2004)https://www.wendangku.net/doc/fa7648139.html,ndscape metrics

Landscape patterns of bird diversity conservation areas were calculated using FRAGSTATS in the GIS software ArcView 3.2(McGarigal and Marks,1995).To eliminate redundant land-use

C.-F.Wu et al./Journal of Environmental Management 92(2011)1809e 1820

1812

pattern data,eight landscape indices,namely,the number of patches (NP),mean patch size(MPS),mean shape index(MSI),mean nearest-neighbor distance(MNN),mean proximity index(MPI),total core area (TCA),largest patch index(LPI),and largest shape index(LSI)were applied to capture the con?guration and connectivity of the conser-vation areas.To evaluate changes in land-use patterns under the various bird diversity conservation scenarios,six landscape indices, namely,the number of patches(NP),mean patch size(MPS),mean shape index(MSI),total core area(TCA),mean core area(MCA)and mean nearest neighbor index(MNN),were used to specify the con?guration and connectivity of land-use types at the landscape scale.The percentage change,P*,in the landscape indicator was then calculated to compare the changes in landscape patterns under?ve protection scenarios.

P*?

V xiàV x1

100%;(3)

where P*is the percentage change in the landscape index between the baseline year and the simulation year;V x1is the landscape index value for the baseline year,and V x i is the landscape index value for the simulation year.

2.6.Conservation of bird diversity scenarios

For each land-use type,conversion rules were adopted to specify the conditions under which the land use may change in the next time step.The rules are derived from the landscape restoration and sub-watershed management policy adopted by the Soil and Water Conservation Bureau of the Council of Agriculture in2006.One rule speci?es that built-up areas,agricultural land,areas of betel nut cultivation,forested area,and grasslands can be converted to any of the?ve land-use types,but water bodies and landslide areas cannot be converted to any other land use.In addition?ve bird diversity conservation policies are generated based on the?ndings of the bird diversity area.The?rst(baseline)policy is a management plan for protecting the diversity of bird species when the predicted bird diversity value exceeds 3.6,which is the smallest value for a conservation area.The second to?fth protection policies gradually increase the size of the protected area and reduce the bird diversity values in the following order3.2,2.8,2.4and2.0.Based on these spatial policies and conversion rules,?ve scenarios were established. Scenarios A,B,C,D and E require bird diversity values larger than2.0,

2.4,2.8,

3.2and3.6respectively.

3.Results

3.1.Bird diversity and its relationship to land-use and landscape factors

Eighty-four bird species were identi?ed in the study area.More speci?cally,56and67species were identi?ed,respectively,in autumn and winter2006;and68and76were identi?ed,respec-tively,in the spring and summer2007.Resident,winter and non-native birds represented64.3%,9.5%and1.2%of the total pop-ulation respectively.Bird species with high populations included Apus af?nis,Pericrocotus solaris,Passer montanus,Lonchura striata, Lonchura punctulata,Alcippe morrisonia,Pycnonotus sinensis,Hyp-sipetes madagascariensis and Paradoxornis webbianus.Apus af?nis is normally found close to residential areas.Lonchura striata,Lonchura punctulata,Alcippe morrisonia and Paradoxornis webbianus roost in agricultural areas and grasslands.Pericrocotus solaris and Hyp-sipetes madagascariensis are the major roosting species in forested areas.Species like Accipiter soloensis and Lanius cristatus,which visit in the winter,were found in the Shangan sub-watershed.The spatial distribution of bird species was not uniform across the study area(Fig.2),but was concentrated mainly in the central and western areas.Sites C1(2.97),D5(2.27),D7(2.69),A3(2.84)and A2 (2.80)had the highest mean Shannon and Weaver diversity index values in all four seasons.Along the main road in the eastern area, sites B5(0.88),E6(0.98)and E9(1.09)had the fewest species and the lowest diversity index values.

Table2lists the results of multiple regression analyses of the Shannon and Weaver index of bird diversity.The results reveal an accumulated r2?0.509,F?3.67,and p<0.01.Eleven

factors,

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

A1A3A5A7A9B1B3B5B7B9C1C3C5C7C9D1D3D5D7D9E1E3E5E7E9

Survey site

Fig.2.Shannon e Wiener bird diversity index values at each survey station.

Table1

Data source and resolution for Shannon and Weaver index of bird diversity and land-

use logistic regression model.

Variables a Data source Scale Resolution

Dependent variables

Land use digital land-use map base on

aerial photography in2007

1:5000e

Independent variables

Landowner Soil and water conservation bureau1:25000e

Rindex Agricultural research institute1:25000e

Soildepth1:25000e

Soilorg1:25000e

SoilpH1:25000e

SoilK1:25000e

DTM Satellite survey center,Department

of Land Administration

e40?40m

Slope Calculated from DEM e40?40m

Driver Calculated from land-use map in2007e40?40m

Droad e40?40m

D.S.W.C.S.Soil and water conservation bureau e40?40m

a Landowner:land ownership;Rindex:rainfall erosion index;Soildepth:soil

depth;Soilorg:the proportion of organic matter in the s oil;SoilpH:soil pH;SoilK:

soil erosion coef?cient;DTM:altitude;slope:Slope;Driver:distance to rivers;

Droad:distance to roads;D.S.W.C.S.:distance to soil and water conservation areas.

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e18201813

including particular land-use types,had a signi ?cant and positive effect on bird diversity.They included water bodies (1.774),land-slide areas (3.255),agricultural land (1.818)and betel nut cultiva-tion (1.125).Edge density (1.41),slope (0.02),distance to built-up area (0.001),and mean shape index (0.003)were positively related to bird diversity;while the distance to roads (à0.001),number of patches (à0.128)and mean patch edge (à0.004)were negatively related to bird diversity.The distribution of bird species diversity was predicted using the above multiple regression model for mapping the Shannon and Weaver diversity based on 2007data.Fig.3shows the map of predicted bird diversity.In this study,the area with the greatest diversity was along the Erbukeng,Sanbu-keng,Liulongkeng and Chenyouland streams.The eastern area is classi ?ed as broadleaf evergreen subtropical forest in which the most abundant plants are Cyclobalanopsis longinux ,Cyclobalanopsis gilva and Lauraceae.This area is associated with low bird species diversity,whereas agricultural areas and areas of betel nut culti-vation in the middle and western parts of the Shangan sub-watershed are associated with high bird species https://www.wendangku.net/doc/fa7648139.html,ndscape patterns under various sub-watershed management policies

A bird species diversity map was used to identify the area of bird diversity protection.Five protection scenarios (A,B,C,D and E),

associated with Shannon and Weaver diversity values larger than 2.02.4,2.8,3.2and 3.6respectively,were constructed.Eight landscape indices,namely,the number of patches (NP),mean patch size (MPS),mean shape index (MSI),mean nearest-neighbor distance (MNN),mean proximity index (MPI),total core area (TCA),largest patch index (LPI),and largest shape index (MSI),were used to specify the con ?guration and connectivity of the conservation area in FRAG-STATS in the GIS software ArcView 3.2(McGarigal and Marks,1995).Fig.4plots the relationships between the landscape indices and the ?ve Shannon and Weaver diversity values based on the regression model by Microsoft Excel.Five regression models,exponential,linear,polynomial,log-linear and power,will be ?tted for each landscape indices and ?ve Shannon and Weaver diversity values.The model shows the highest r 2value is the best model to illustrate the relationship between landscape index and Shannon and Weaver diversity values.Conclusions of regression models shown that number of patches (NP)declined from scenario A to E (Fig.4a),but the mean nearest-neighbor distance index (MNN)increased (Fig.4e).Fig.4b,c,g and h shows,respectively,that the mean patch size (MPS),mean shape index (MSI),largest patch index (LPI),and largest shape index (LSI)declined from scenario A to E.Both the mean proximity index (MPI)and the total core area (TCA)declined exponentially from scenario E to A (Fig.4d and f).

3.3.Simulated land-use patterns and bird diversity under various conservation scenarios

Table 3shows the estimated coef ?cients and Relative Operating Characteristic (ROC)values obtained using the forward stepwise logistic regression model for all land-use types.The ?tted logistic models were used to calculate the occurrence probabilities for all land-use types.The ROC values for the models ranged from 0.69to 1.0,suggesting that the models explained the spatial variation in land use adequately.

Land ownership,soil depth,amount of organic matter in the soil,soil erosion coef ?cient,altitude and slope were positively related to the probability that land was forested,while the rainfall erosion index,soil pH,distance to roads,and distance to soil and water conservation engineering sites were negatively related to the pre-dicted probability that land was forested.Factors that are positively related to using land for the betel nut cultivation include rainfall erosion index,soil depth,soil pH,slope and distance to https://www.wendangku.net/doc/fa7648139.html,nd ownership,the amount of organic matter in the soil,the soil erosion coef ?cient,altitude,the distance to roads,and the distance to soil and water conservation engineering sites were negatively related to the occurrence of betel nut cultivation.

Factors that are negatively related to the agricultural land include land ownership,soil depth,the amount of organic matter in the soil,slope,the distance to roads,and the distance to soil and conservation engineering sites.Moreover,the rainfall erosion index,soil pH,soil erosion coef ?cient,altitude and distance to river are positively related to the presence of agricultural land.Factors that are negatively related to grassland include the rainfall erosion index,soil erosion coef ?cient,altitude,slope,distance to roads,and distance to soil and water conservation engineering https://www.wendangku.net/doc/fa7648139.html,nd ownership,soil depth and distance to river are positively related to the presence of agricultural land.

Factors that are negatively related to grassland include the rainfall erosion index,soil erosion coef ?cient,altitude,slope,distance to roads,and distance to soil and water conservation engineering https://www.wendangku.net/doc/fa7648139.html,nd ownership,soil depth and distance to rivers are positively related to the presence of agricultural land.Factors that are positively related to built-up land include the rainfall erosion index,soil depth,and distance to soil and water conser-vation engineering https://www.wendangku.net/doc/fa7648139.html,nd ownership,soil pH,soil erosion

Table 2

Multiple regression analysis a of the Shannon and Weaver index of bird diversity.Variables

Regression coef ?cient (?standard error)Signi ?cance (p )Percentage of landslide areas 3.255?1.5580.043Percentage of agriculture areas 1.818?0.442<0.0001Percentage of water areas 1.774?0.6790.013Percentage of betel nut areas 1.125?0.4460.016Number of patch à0.128?0.0680.067Edge density

1.41?0.5480.014Mean shape index 0.003?0.0010.004Mean patch edge

à0.004?0.00010.007Distance to built-up areas 0.001?0.0010.077Distance to roads à0.001?0.00040.051Slope 0.02?0.0090.037Constant

1.3?

0.772

0.081

R value ?0.713,r 2?

0.509.

Fig.3.Predictive distribution of bird species diversity in 2007.

C.-F.Wu et al./Journal of Environmental Management 92(2011)1809e 1820

1814

coef ?cient,slope,distance to rivers and distance to roads are negatively related to the presence of built-up land.

Factors that are negatively related to the presence of water bodies include land ownership,rainfall erosion index,soil depth and distance to rivers;however,the distance to soil and water conservation engineering sites is positively related to the presence of water bodies.Factors that are positively related to landslide land include the soil erosion coef ?cient,altitude and slope.Distance to rivers and distance to roads are negatively related to the occurrence of landslides.

Based on the logistic regression models,land-use demand,spatial policies and land-use conversion rules,the CLUE-s

model

https://www.wendangku.net/doc/fa7648139.html,ndscape metrics of bird diversity area conservation scenarios at landscape level (a)Number of Patches,(b)Mean Patch Size,(c)Mean Shape Index,(d)Mean Proximity Index,(e)Mean Nearest-Neighbor Distance,(f)Total Core Area (TCA),(g)Largest Patch Index (LPI),(h)Largest Shape Index (MSI).

Table 3

Logistics regression model for land-use types.Variable b Forest Betel nut Agriculture Grassland Built-up Water Landslide Landowner 2.178à0.357à1.167 3.036à2.662à0.700e a Rindex à0.0150.0110.019à0.0020.009à0.026e Soildepth 0.0050.006à0.0070.0090.006à0.016e Soilorg 0.012à0.005à0.011e e

e e SoilpH à0.5860.5890.192e

à0.093e e

SoilK 0.033à0.0460.021à0.007à0.018e 0.604DTM 0.003à0.0060.0003à0.001e

e 0.008Slope 0.0840.007à0.034à0.012à0.083e

0.011Driver e

0.0020.00070.003à0.0006à0.766à0.007Droad à0.002à0.0007à0.002à0.0006à0.020e

à0.007D.S.W.C.S.à0.0001à0.0005à0.0003à0.0010.00040.001e

Constant 4.103à6.899à14.503à2.574à4.43124.889à28.251ROC

0.88

0.73

0.71

0.69

0.93

0.91

a not signi ?cant and not included in model at 0.05signi ?cant level.

Landowner:land ownership;Rindex:rainfall erosion index;Soildepth:soil depth;Soilorg:the proportion of organic matter in the soil;SoilpH:soil pH;SoilK:soil erosion coef ?cient;DTM:altitude;slope:Slope;Driver:distance to rivers;Droad:distance to roads;D.S.W.C.S.:distance to soil and water conservation areas.

C.-F.Wu et al./Journal of Environmental Management 92(2011)1809e 18201815

was applied to simulate land-use changes under the?ve bird diversity protection scenarios for the period2008e2017.Fig.5 shows the resulting map associated with each land-use change between2007and2017under each scenario.The map shows that the areas that will change most frequently are located in the eastern and southwestern parts of the Shangan watershed.In the eastern part,land use will change from grassland,landslide areas and cultivated land to forest.In the southwest area,cultivated land will be converted to built-up land and land for betel nut cultivation.

Six landscape indices were calculated using CLUE-s simulation maps for the period2007e2017.Fig.6shows the percentage changes of the six indices during the study period.The number of patches(NP)will increase under all?ve scenarios from2007to 2017(Fig.6a).The largest increase will be under Scenario A (11.34%),and the smallest will be under scenario B(7.92%).The increases in scenarios C,D and E are expected to be9.5%,10.0%and 9.26%respectively.Fig.6b shows that,in all scenarios,the mean patch size(MPS)will decrease from2007to2017.The7.6%decline in scenario B will be the smallest and the10.4%decline in scenario A will be the largest.The rates of decline in scenarios C,D and E will be8.88%,8.88%and8.82%respectively.The mean shape index(MSI) will fall in all?ve scenarios from2007to2017(Fig.6c).The decline in Scenario A(3.82%)will be the largest,and decline in Scenario B (3.18%)will be the smallest.The MSI declines in Scenarios C,D and E will be similar(3.21%).

Fig.6d shows that the total core area(TCA)will increase under all scenarios between2007and2017.The increase in scenario B, 0.568%,will be the largest and that in scenario A,0.467%,will be the smallest.The increases under scenarios C,D and E will be0.554%, 0.475%and0.488%respectively.The mean core area(MCA)will decline under all scenarios(Fig.6e).Scenario B will decline the least (0.578%),and scenario D will decline the most(2.874%).The declines in scenarios A,C and E will be1.705%,1.744%and2.299% respectively.Finally,the mean nearest-neighbor distance index (MNN)will increase in all scenarios(Fig.6f).The increase in scenario A(4.14%)will be the largest,while that in scenario D

will

https://www.wendangku.net/doc/fa7648139.html,nd use changes between land use in2007and(a)conservation scenario A,(b)conservation scenario B,(c)conservation scenario C,(d)conservation scenario D,(e) conservation scenario E in2017.

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e1820

1816

be the smallest.The increases in scenarios B,C and E will be 0.96%,1.0%and 2.88%respectively.

Fig.7plots the total values of the diversity indices calculated by the Shannon and Weaver diversity regression model.Scenario B is associated with the largest total suitability value (135,343.14),while scenario E has the lowest value (135,080.02).The values for scenarios A,C and D are 135,324.54,135,335.5and 135,152.89respectively.4.Discussion

4.1.Factors that affect bird diversity

In the ?fty samples,the number of birds ranged from 10to 429,and the mean Shannon and Weaver diversity index ranged from 0.83to 2.97over the study watershed.Eighty-four bird species were recorded in the area.At the three most species-rich survey locations (hotspots),the numbers of bird species varied between 29and 41;while in the three least species-rich locations (coldspots),there were only 3to 4species.The most common species were Passer montanus ,with populations of 8and 72at sites E4and C4respectively;Lon-chura striata with populations of 59and 58at sites A3and C1respectively;Apus af ?nis with populations of 78and 67at sites A4and A3respectively;and Alcippe morrisonia with a population of 76

at site C1.The four most common bird species represented 36.24%of all the birds observed.

Landscape features,such as hedgerows,untilled land,bushes,trees and anthropogenic land use,have a marked effect on the abundance of a bird species (Scozzafava and De Sanctis,2006).In forested landscapes,the most important explanatory factors for variations in the relative abundance of bird species in forest fragments are:(1)species abundance within continuous forest,which is expected based on the similarity of forest composition and

structure;

Fig.6.Percent of landscape metrics change between 2007and 2017.Number of Patches (NP),(b)Mean Patch Size (MPS),(c)Mean Shape Index (MSI),(d)Total Core Area (TCA),(e)Mean Core Area (MCA),(f)Mean Nearest-Neighbor Distance

(MNN).

Fig.7.Sum values of predictive Shannon and Weaver ’s diversity.

C.-F.Wu et al./Journal of Environmental Management 92(2011)1809e 18201817

(2)presence in the matrix;(3)migratory strategy;and(4)habitat association(Renjifo,2001).Moderate levels of development may increase the overall species diversity,but reduce native bird diversity; while intensive development reduces both the overall and the native species diversity(Blair,1996).The eastern area of the Shangan sub-watershed is characterized by high elevation,steep slopes,and forest cover with a few plant species.The bird species Pericrocotus solaris and Hypsipetes madagascariensis inhabit this area.The central and western parts of the study area are characterized by low elevation, gentle slopes and a variety of land uses(cultivated land,water bodies, grassland and betel nut cultivation).The composition and con?gu-ration of habitats in the central and western areas differ greatly from the almost natural habitat in the eastern area.Agricultural?elds, grassland,?sh in rivers or ponds and the seeds of forest plants are the major food sources for birds in the study area.Our survey results show that61.9%of the species are not major roosting species.Passer mon-tanus,Pycnonotus sinensis,Lonchura striata,Lonchura punctulata, Alcippe morrisonia and Paradoxornis webbianus are the major roosting species.

Many studies have developed models that relate the distribution of bird species to environmental variables(Buckland and Elston,1993; Bustamante,1997;Tucker et al.,1997;Siriwardena et al.,1998;Saab, 1999;Luoto et al.,2004).In this work,the Shannon and Weaver diversity regression model shows that land-use diversity increases bird species diversity.The latter is typically the greatest in areas where the major land uses are cultivated land,water bodies,and betel nut cultivation,and where the land is prone to landslides,close to a road,far from a built-up area and on a steep slope.Moreover,a small number of irregularly shaped land-use patches are highly correlated with bird species diversity.Some researchers have suggested that abiotic environmental variables,as well as habitat structure and composition,could probably be used to improve the accuracy of bird habitat descriptions(Dettmers and Bart,1999;Luoto et al.,2004). Such models can also reveal changes in the distribution of species over time,or be used to evaluate the suitability of an area for sup-porting bird diversity if the area has not been studied directly (Buckland and Elston,1993).In this work,the r2value for the habitat suitability regression model was0.509,indicating that the multiple regression model can predict bird diversity accurately.

https://www.wendangku.net/doc/fa7648139.html,ndscape patterns under various bird diversity conservation scenarios

Numerous studies have demonstrated the relationships between the diversity of animal species and the properties of habitat patches, including the size,shape and diversity of habitats(e.g.,Bolger et al., 1997;Deacon and Mac Nally,1998).High connectivity of habitats also favors larger local populations through immigration(Renjifo, 2001).Variations in bird abundances and biodiversity among land-scapes may be associated with variations in?oristic composition and structure(Renjifo,2001;Herrando and Brotons,2002;Fairbanks, 2004;Mitchell et al.,2006;Ernoult et al.,2006).In this study,land-scape metrics were applied to analyze the landscape patterns under the?ve scenarios of protected bird diversity.The values of the number of patches(NP),mean patch size(MPS),mean shape index(MSI), mean proximity index(MPI),largest patch index(LPI),largest shape index(LSI)and total core area(TCA)in the landscape of the protected scenario are higher than those in the landscapes of the other scenarios.However,the nearest-neighbor distance index(MNN)in the landscape of the protected scenario is lower than that in the landscapes of the other scenarios.The above landscape metric results show that the protected area is comprised of some large patches, a large total core area,irregular patch shapes,and high connectivity. Additionally,the r2values vary from0.930to0.993if eight landscape indices are used,indicating that the regression models can explain indicator changes adequately.

4.3.Policies for conserving bird diversity and changes in land-use

One of the main goals of nature conservation and land-use planning is to identify and preserve areas that promote biodiversity (Luoto et al.,2004).Unfortunately,data on spatial biodiversity patterns,and particularly the distribution and richness of species in landscape mosaics,is often sparse or simply unavailable;moreover, it is usually expensive to acquire(Margules and Austin,1991;Scott et al.,1993;Gaston,1996;Luoto et al.,2004).In this work,land-use changes from2007to2017were simulated using the CLUE-s model,in which the relationships between land-use types and their driving factors are evaluated by stepwise logistic regression (Verburg et al.,2002;Lin et al.,2009).The logistic model’s results for all types of land-use show that the locations of forested areas are determined by both socio-economic parameters(land owner-ship,distance to roads,and distance to soil and water conservation structures)and biophysical characteristics,such as the rainfall erosion index,soil depth,soil organic matter,soil pH,soil erosion coef?cient,altitude and slope.These analytical results indicate that forest locations in the Shangan watershed are affected by both physical and socio-economic characteristics,and particularly by land ownership and soil conditions.The locations and distribution of betel nut crops are governed by all of the above-mentioned biophysical and socio-economic factors.The locations of agricul-tural land are determined by all of the listed biophysical and socio-economic characteristics,and particularly by land ownership and soil pH.All factors,except the proportion of organic matter in the soil and the soil pH balance,affect the location of grasslands.

Moreover,all factors,except the proportion of organic matter in the soil and altitude affect the locations of built-up areas.The slope and distance to roads were negative factors in the use of land for building,indicating that built-up areas expand along roads in areas with low slopes.The locations of landslide areas were determined by biophysical parameters(altitude,slope,and soil erosion coef?-cients,distance to river)as well as socio-economic characteristics, such as distance to major roads.These analytical results reveal that a complex mix of physical and socio-economic factors affect land-slide areas in the examined watershed.

Biophysical parameters such as altitude,slope,and soil erosion coef?cients,as well as the distance to rivers affected the distribu-tions of most types of land use in the study area.The parameters affect biophysical processes that result in speci?c land-use patterns (Lambin,1994;Serneels and Lambin,2001).The logistic regression results also verify that socio-economic factors determine the potential advantages of a particular type of land use in a certain location,suggesting that watershed management affects land use in the watershed of interest.Finally,in this work,the ROC values varied between0.69and1.00,depending on the land-use type, suggesting that the logistic regression model explained the land-use distribution effectively.

4.4.Effects of policies for protecting bird diversity on bird diversity and landscape patterns

Spatial models,such as explicit models for forecasting types of land use,can help planners evaluate the long-term effects of patterns of development on the landscape and the value of that development(Wear and Bolstad,1998;Turner et al.,2001;Lin et al., 2007,2009).In this study,the CLUE-s model,which combines land-use demands and policies to protect bird diversity,was used to simulate land-use scenarios successfully in the examined watershed during the speci?ed period.The simulation results demonstrate that

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e1820 1818

changes in land-use occur more frequently in the high and heavily populated eastern and southwestern areas of the Shangan sub-watershed than elsewhere(Fig.5).

Landscape metrics are critical to landscape planning because they reveal changes in landscape patterns(Corry and Nassauer, 2005;Lin et al.,2007,2009).Bennett et al.(2006)noted that animals are affected by the following four aspects of the spatial con?guration of land use:subdivided mosaics,the size of habitat patches,the structural connectivity of a mosaic,and the lengths of the edges in the mosaic.Herrando and Brotons(2002)found that the abundance of forest birds in Mediterranean areas was positively related to the patch size and the irregular shape of a forest. Furthermore,a study of Dutch landscapes found that the richness of bird species in heathland declined as habitats became dis-aggregated(Olff and Ritchie,2002).In the present study,the landscape patterns were similar at the landscape scale in all scenarios;however,bird diversity conservation policy scenario B involved larger patches,more irregular shapes,a larger total core area,and more closely positioned patches than the other scenarios. Conservation policy scenario A involved more isolated types of land-use,smaller core areas,and fewer irregularly shaped areas than the other four scenarios.The results demonstrate that larger areas of bird diversity can be protected by increasing the patch size, the habitat core area,the lengths of edges,and the connectivity of habitat patches.

Species diversity maps are important for conservation planning (Gould,2000)as well as for identifying areas that are particularly important for high biodiversity,where conservation resources must be concentrated(Luoto et al.,2004).We compiled a biodi-versity distribution map and calculated the Shannon and Weaver diversity values for conservation scenarios,which are critical to developing effective protection policies.The Shannon and Weaver diversity regression model was used to calculate the sum of the diversity values in the study area for?ve conservation scenarios(Fig.7).Scenarios A,B and C have higher total diversity values than scenarios D or E.The total diversity value under scenario B was135,343.14,which was the highest of all scenarios. The results show that protecting large areas of bird species diversity is effective in increasing bird diversity in the studied sub-watershed.Our analysis of bird diversity indicates that larger conservation areas provide more suitable bird habitats,particu-larly in scenario B.

5.Conclusion

Land use development,disturbances and urbanization present particular challenges in the area of biodiversity conservation.The management of land use in the Shangan watershed will certainly become an increasingly important aspect of nature conservation policies because of the spatial variability of biodiversity in the mountainous regions of central Taiwan.The unique land-use structures and habitats in this area protect bird species diversity and may provide seasonal habitats that enhance total biodiversity, especially in highly disturbed areas.This study combined bird survey data,regression modeling,land-use modeling and landscape metrics to simulate and evaluate bird diversity conservation policies and their effect on landscape patterns and bird diversity in a speci?c area.Bird species are highly concentrated in the central and western parts of the study area.Moreover,land-use diversity increases bird diversity.The results obtained by applying the land-use simulation model to the Shangan sub-watershed indicate that the eastern and southwestern areas will change most frequently between2007and 2017.Moreover,according to the model,increasing the size of areas to protect bird diversity could increase the patch size,habitat core area,edge effect and habitat connectivity.The Shannon e Weaver diversity regression model demonstrates that protecting bird species diversity in large areas would also be an effective way to increase bird diversity in the study area.The proposed modeling approach provides an effective tool and yields useful information for watershed ecological planning as well as for land-use planning, management,and policymaking.Future studies could monitor bird species to validate the modeling results and adapt them for management purposes.Moreover,scale can impact signi?cantly on the measurement and quantitative description of land-use patterns,in?uences the behavior of model parameters used to describe land-use change processes.Thus,future work should also focus on the scale effects on land-use and landscape pattern change in sub-watershed area.

Acknowledgments

The authors would like to thank the National Science Council of the Republic of China,Taiwan,for?nancially supporting this research under contract nos.NSC97-2410-H-167-002-.The author would also like to appreciate Dr.Verburg and the CLUE-S group for providing the CLUE-S model.

References

Bennett,A.F.,Hinsley,S.A.,Bellamy,P.E.,Swetnam,R.D.,Mac Nally,R.,2004.Do regional gradients in land-use in?uence richness,composition and turnover of bird assemblage in small woods?Biological Conservation119,191e206. Bennett,A.F.,Radford,J.Q.,Haslem,A.,2006.Properties of land mosaics:implica-tions for nature conservation in agricultural environments.Biological Conser-vation133,250e264.

Blair,R.B.,https://www.wendangku.net/doc/fa7648139.html,nd use and avian species diversity along an urban gradient.

Ecological Applications6,506e512.

Bolger, D.T.,Scott,T.A.,Rotenberry,J.T.,1997.Breeding bird abundance in an urbanizing landscape in coastal southern California.Biological Conservation11, 406e421.

Buckland,S.T.,Elston,D.A.,1993.Empirical models for the spatial distribution of wildlife.Journal of Applied Ecology30,478e495.

Buckland,S.T.,Anderson,D.R.,Burnham,K.P.,Laake,J.L.,Borchers,D.L.,Thomas,L., 2001.Introduction to Distance Sampling:Estimating Abundance of Biological Populations.Oxford University,U.K.,p.432.

Bustamante,J.,1997.Predictive models for lesser kestrel Falconaumanni distribu-tion,abundance and extinction in Southern Spain.Biological Conservation80, 153e160.

Castella,J.C.,Pheng Kam,S.,Dinh Quang,D.,Verburg,P.H.,Thai Hoanh,C.,2007.

Combining topdown and bottom-up modelling approaches of land-use/cover change to support public policies:application to sustainable management of natural resources in northern https://www.wendangku.net/doc/fa7648139.html,nd-use Policy24(3),531e545. Chapman,K.A.,Reich,P.B.,https://www.wendangku.net/doc/fa7648139.html,nd use and habitat gradients determine bird community diversity and abundance in suburban,rural and reserve landscapes of Minnesota,USA.Biological Conservation135,527e541.

Corry,R.C.,Nassauer,J.I.,2005.Limitations of using landscape pattern indices to evaluate the ecological consequences of alternative plans and https://www.wendangku.net/doc/fa7648139.html,nd-scape and Urban Planning72,265e280.

Deacon,J.N.,Mac Nally,R.,1998.Local extinction and nestedness of small-mammal faunas in fragmented forest of central Victoria,Australia.Paci?c Conservation Biology4,122e131.

Dettmers,R.,Bart,J.,1999.A GIS modeling method applied to predicting forest songbird habitat.Ecological Applications9,152e163.

Devictor,V.,Jiguet, F.,https://www.wendangku.net/doc/fa7648139.html,munity richness and stability in agricultural landscapes:the importance of surrounding habitats.Agriculture Ecosystems and Environment120,179e184.

Ernoult,A.,Tremauville,Y.,Cellier,D.,Margerie,P.,Langlois,E.,Alard,D.,2006.

Potential landscape drivers of biodiversity components in a?ood plain:past or present patterns?Biological Conservation127,1e17.

Fairbanks,D.H.K.,2004.Regional land-use impacts affecting avian richness patterns in Southern Africa-insights from historical avian atlas data.Agriculture Ecosystems and Environment101,269e288.

Felton,M.,1996.Natura2000,the ecological network of the European union:using buffer areas and corridors to reinforce areas designated by member states.In: Nowicki,P.,Bennett,G.,Middleton, D.(Eds.),Perspectives in Ecological Networks.ECNC,Arnhem,Netherlands,pp.133e141.

Fernández-Juricic,E.,2004.Spatial and temporal analysis of the distribution of forest specialists in an urban-fragmented landscape(Madrid,Spain)implica-tions for local and regional bird https://www.wendangku.net/doc/fa7648139.html,ndscape and Urban Planning 69,17e32.

Forman,R.T.T.,1995.Some general principles of landscape and regional ecology.

Landscape Ecology10,133e142.

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e18201819

Forman,R.T.T.,Godron,M.,https://www.wendangku.net/doc/fa7648139.html,ndscape Ecology.Wiley,New York.

Goetz,S.,Steinberg, D.,Dubayah,R.,Blair, B.,https://www.wendangku.net/doc/fa7648139.html,ser remote sensing of

canopy habitat heterogeneity as a predictor of bird species richness in an

eastern temperate forest,USA.Remote Sensing of Environment108,

254e263.

Gaston,K.J.,1996.Biodiversity:a Biology of Numbers and Difference.Blackwell

Scienti?c,Oxford.

Gould,W.A.,2000.Remote sensing of vegetation,plant species richness,and

regional diversity hotspots.Ecological Applications10,1861e1870.

Graf,R.F.,Bollmann,K.,Suter,W.,Bugmann,H.,2005.The importance of spatial

scale in habitat models:capercaillie in the Swiss https://www.wendangku.net/doc/fa7648139.html,ndscape Ecology20,

703e717.

Herrando,S.,Brotons,L.,2002.Forest bird diversity in Mediterranean areas affected

by wild?res:a multi-scale approach.Ecography25,161e172.

Holzk?mper,A.,Lausch,A.,Seppelt,R.,2006.Optimizing landscape con?guration to

enhance habitat suitability for species with contrasting habitat requirements.

Ecological Modelling198,277e292.

Lambin,E.F.,1994.Modeling Deforestation Processes:A Review.Tropical Ecosystem

Environment Observations by Satellites(TREES).Series B:Research Report no.1.

European Commission,Luxembourg.

Lambin, E.F.,Turner II, B.L.,Geist,H.J.,Agbola,S.B.,Angelsen, A.,Bruce,J.W.,

Coomes,O.,Dirzo,R.,Fischer,G.,Folke, C.,George,P.S.,Homewood,K.,

Imbernon,J.,Leemans,R.,Li,X.,Moran,E.F.,Mortimore,M.,Ramakrishnan,P.S.,

Richards,J.F.,Sk anes,H.,Steffen,W.,Stone,G.,Svedin,U.,Veldkamp,T.A.,

Vogel,C.,Xu,J.,2001.The causes of land-use and landcover change:moving

beyond the myths.Global Environmental Change11(4),261e269.

Langanke,T.,Rossner,G.,Vr s c aj, B.,Lang,S.,Mitchley,J.,2005.Selection and application of spatial indicators for nature conservation at different institu-

tional levels.Journal for Nature Conservation13,101e114.

Lin,Y.P.,Chang,T.K.,Wu,C.F.,Chiang,T.C.,Lin,S.H.,2006.Assessing impacts of

typhoons and the Chi-Chi earthquake on Chenyulan watershed landscape

pattern in central Taiwan using landscape metrics.Environmental Management

38(1),108e125.

Lin,Y.P.,Hong,N.M.,Wu,P.J.,Wu,C.F.,Verburg,P.H.,2007.Impacts of land use

change scenarios on hydrology and land use patterns in the Wu-Tu watershed

in Northern https://www.wendangku.net/doc/fa7648139.html,ndscape and Urban Planning80,111e126.

Lin,Y.P.,Yeh,M.S.,Deng,D.P.,Wang,Y.C.,2008a.Geostatistical approaches and

optimal additional sampling schemes for spatial patterns and future sampling

of bird diversity.Global Ecology and Biogeography17,175e188.

Lin,Y.P.,Wu,P.J.,Hong,N.M.,2008b.The effects of changing the resolution of land-

use modeling on simulations of land-use patterns and hydrology for a water-

shed land-use planning assessment in Wu-tu,https://www.wendangku.net/doc/fa7648139.html,ndscape and Urban

Planning87,54e66.

Lin,Y.P.,Verburg,P.H.,Chang,C.R.,Chen,H.Y.,Cheng,M.H.,2009.Developing and

comparing optimal and empirical land-use models for the development of an

urbanized watershed forest in https://www.wendangku.net/doc/fa7648139.html,ndscape and Urban Planning92,

242e254.

Luoto,M.,Virkkala,R.,Heikkinen,R.K.,Rainio,K.,2004.Predicting bird species

richness using remote sensing in boreal agricultural-forest mosaics.Ecological

Applications14(6),1946e1962.

Magurran, A.E.,1988.Ecological Diversity and Its Measurement.Croom Helm,

London.

Margules,C.R.,Austin,M.P.,1991.Nature Conservation:Cost Effective Biological

Surveys and Data Analysis.CSIRO,Melbourne,Australia.

McGarigal,K.,Marks,B.J.,1995.FRAGSTATS:Spatial Pattern Analysis Program for

Quantifying Landscape Structure USDA Forest Service General Technical Report.

PNW-351.

Miller,J.N.,Brooks,R.P.,Croonquist,M.J.,1997.Effects of landscape patterns on biotic

https://www.wendangku.net/doc/fa7648139.html,ndscape Ecology12,137e153.

Mitchell,M.S.,Rutzmoser,S.H.,Wigley,T.B.,Loehle,C.,Gerwin,J.A.,Keyser,P.D.,

Lancia,R.A.,Perry,R.W.,Reynolds,C.J.,Thill,R.E.,Weih,R.,White,D.,Wood,P.B.,

2006.Relationships between avian richness and landscape structure at multiple

scales using multiple landscapes.Forest Ecology and Management221,

155e169.

Mitchell,M.S.,Reynolds-Hogland,M.J.,Smith,M.L.,Wood,P.B.,Beebe,J.A.,

Keyser,P.D.,Loehle,C.,Reynolds,C.J.,Deusen,P.V.,D.W.Jr.,,2008.Projected

long-term response of Southeastern birds to forest management.Forest Ecology

and Management256,1884e1896.

M?rtberga,U.M.,Balforsa,B.,Knolb,W.C.,https://www.wendangku.net/doc/fa7648139.html,ndscape ecological assessment:

a tool for integrating biodiversity issues in strategic environmental assessment

and planning.Journal of Environmental Management82,45e470.National Land Surveying and Mapping Center,2006.Work criterion of1/1000 digital topographic map.Technical report National Land Surveying and Mapping Center in Taiwan.

Oja,T.,Alamets,K.,P?rnamets,H.,2005.Modelling bird habitat suitability based on landscape parameters at different scales.Ecological Indicators5,314e321. Olff,H.,Ritchie,M.E.,2002.Fragmented nature:consequences for biodiversity.

Landscape and Urban Planning58,83e92.

Overmars,K.,Verburg,P.,2005.Analysis of land use drivers at the watershed and household level:linking two paradigms at the Philippine forest fringe.Inter-national Journal of Geographical Information Science19,125e152.

Pino,J.,Rodaáa,F.,Ribasb,J.,Pons,X.,https://www.wendangku.net/doc/fa7648139.html,ndscape structure and bird species richness:implications for conservation in rural areas between natural parks.

Landscape and Urban Planning49,35e48.

Rafe,R.W.,Usher,M.B.,Jefferson,R.G.,1985.Birds on reserves:the inuence of area and habitat on species richness.Journal of Applied Ecology22,327e335. Renjifo,L.M.,2001.Effect of natural and anthropogenic landscape matrices on the abundance of subandean bird species.Ecological Applications11,14e31. Saab,V.,1999.Importance of spatial scale to habitat use by breeding birds in riparian forests:a hierarchical analysis.Ecological Applications9,135e151. Scott,J.M.,Davis,F.,Csuti,B.,Noss,R.,Butter?eld,B.,Groves,C.,Anderson,H., Caicco,S.,D’Erchia,F.,Edwards,T.C.,Ulliman,J.J.,Wright,R.G.,1993.Gap anal-ysis:a geographic approach to protection of biodiversity.Supplement Journal of Wildlife Management123,57.Wildlife Monographs.

Scozzafava,S.,De Sanctis,A.,2006.Exploring the effects of land abandonment on habitat structures and on habitat suitability for three passerine species in

a highland area of Central https://www.wendangku.net/doc/fa7648139.html,ndscape and Urban Planning75,23e33. Serneels,S.,Lambin, E.F.,2001.Proximate causes of land-use change in Narok district,Kenya:a spatial statistical model.Agriculture Ecosystems and Envi-ronment85,65e81.

Siriwardena,G.M.,Baillie,S.R.,Buckland,S.T.,Fewster,R.M.,Marchant,J.H., Wilson,J.D.,1998.Trends in the abundance of farmland birds:a quantitative comparison of smoothed common birds census indices.Journal of Applied Ecology35,24e43.

Smith,C.M.,Wachob,D.G.,2006.Trends associated with residential development in riparian breeding bird habitat along the snake river in Jackson hole,WY,USA: implications for conservation planning.Biological Conservation128,431e446. Soil and Water Conservation Bureau,https://www.wendangku.net/doc/fa7648139.html,ndscape Ecological Survey and Assess Techniques Applied to Watershed Conservation SWCB-97e170.

Tucker,G.M.,Heath,M.F.,Tomialocjc,L.,Grimmett,R.F.A.,1994.Birds in Europe: Their Conservation Status.Birdlife International,Cambridge.

Tucker,K.,Rushton,S.P.,Sanderson,R.A.,Martin,E.B.,Blaiklock,J.,1997.Modelling bird distributions-a combined GIS and Bayesian rule-based https://www.wendangku.net/doc/fa7648139.html,nd-scape Ecology12,77e93.

Turner,M.G.,Gardner,R.H.,1993.Quantitative methods in landscape ecology:an introduction.In:Turner,M.G.,Gardner,R.H.(Eds.),Quantitative Methods in Landscape Ecology.Springer,New York,pp.3e14.

Turner,M.G.,Gardner,R.H.,O’Neill,R.V.,https://www.wendangku.net/doc/fa7648139.html,ndscape Ecology in Theory and Practice Pattern and Process.Springer e Verlag,New York,p.401.

Verburg,P.H.,Soepboer,W.,Veldkamp, A.,Limpiada,R.,Espaldon,V.,Sharifah Mastura,S.A.,2002.Modeling the spatial dynamics of regional land-use:the CLUE-S model.Environmental Management30(3),391e405.

Verburg,P.H.,Veldkamp,A.,2004.Projecting land use transitions at forest fringes in the Philippines at two spatial https://www.wendangku.net/doc/fa7648139.html,ndscape Ecology19,77e98.

Vesk,P.A.,Nally,R.M.,2006.The clock is ticking-revegetation and habitat for birds and arboreal mammals in rural landscapes of southern Australia.Agriculture Ecosystems and Environment112,356e366.

Vidaurre,W.,Pacheco,L.F.,Roldán,A.I.,https://www.wendangku.net/doc/fa7648139.html,position and abundance of birds in Andean alder(Alnus acuminata)patches with past and present harvest in Bolivia.Biological Conservation132,12e21.

Wear,D.N.,Bolstad,P.,https://www.wendangku.net/doc/fa7648139.html,nd-use changes in southern Appalachian landscape: spatial analysis and forecast evaluation.Ecosystems1,575e594.

White,D.,Minotti,P.G.,Barczak,M.J.,Sifneos,J.C.,Freemark,K.E.,Santelmann,M.V., Steinitz,C.F.,Ross Kiester,A.,Preston,E.M.,1997.Assessing risk to biodiversity from future landscape change.Conservation Biology11,349e360. Woodhouse,S.P.,Good,J.E.G.,Lovett,A.A.,Fuller,R.J.,Dolman,P.M.,2005.Effects of land-use and agricultural management on birds of marginal farmland:a case study in the Llyn peninsula,Wales.Agriculture Ecosystems and Environment 107,331e340.

Yamaura,Y.,Katoh,K.,Fujita,G.,Higuchi,H.,2005.The effect of landscape contexts on wintering bird communities in rural Japan.Forest Ecology and Management 216,187e200.

C.-F.Wu et al./Journal of Environmental Management92(2011)1809e1820 1820