PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Bani, Luciano]On: 12 March 2010Access details: Access Details: [subscription number 919839323]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Italian Journal of ZoologyPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t741771159

Usefulness of coarse grain data on forest management to improve birdabundance modelsD. Massimino a; V. Orioli a; F. Pizzardi a; R. Massa a; L. Bani a

a Dipartimento di Scienze dell'Ambiente e del Territorio, Università degli Studi di Milano-Bicocca,Milano, Italy

First published on: 15 January 2010

To cite this Article Massimino, D., Orioli, V., Pizzardi, F., Massa, R. and Bani, L.(2010) 'Usefulness of coarse grain data onforest management to improve bird abundance models', Italian Journal of Zoology, 77: 1, 71 — 80, First published on: 15January 2010 (iFirst)To link to this Article: DOI: 10.1080/11250000802658058URL: http://dx.doi.org/10.1080/11250000802658058

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Usefulness of coarse grain data on forest management to improve birdabundance models

D. MASSIMINO, V. ORIOLI, F. PIZZARDI, R. MASSA, & L. BANI*

Dipartimento di Scienze dell’Ambiente e del Territorio, Universita degli Studi di Milano-Bicocca, Milano, Italy

AbstractWe tested the use of coarse grain data on forest management systems to improve habitat suitability models in predictingdistribution and abundance of five bird species, nesting in broadleaf forests in Lombardy (Italy). They were the long-tailedtit Aegithalos caudatus, marsh tit Poecile palustris, blue tit Cyanistes caeruleus, wood nuthatch Sitta europaea and short-toedtree-creeper Certhia brachydactyla. Bird data were collected from 4552 point counts from 1992 to 2002. To model thehabitat suitability for each species, we used generalized linear models with negative binomial distribution and log linkfunction. Environmental predictors were land-use fractional covers, forest isolation, elevation, slope, aspect, and coarsegrain data on forest management systems aggregated for administrative provinces. They were selected by a stepwiseprocedure based on the Bayesian Informative Criterion. Predictors on forest management system were useful to improvemodel fit for all cavity-nesting species, which was also shown to be affected by forest isolation. The long-tailed tit, which isnot an obligate cavity nester, did not exhibit any preference with respect either to forest management or isolation.

Keywords: Birds, forest isolation, forestry data, generalized linear models, habitat suitability models

Introduction

Understanding the factors that determine species

distribution and abundance is necessary to undertake

effective conservation actions, for example to contrast

the decline of endangered species, select conservation

priority areas (e.g. Maiorano et al. 2006; Rondinini

et al. 2006; Zielinski et al. 2006; Tapia et al. 2007) or

assess the effects of environmental changes (e.g. Berry

et al. 2002; Thuiller 2003; Araujo et al. 2004; Puglisi

et al. 2005; Seoane et al. 2006). For this reason,

habitat suitability models (HSMs) have been increas-

ingly used in the last decades. HSMs aim to describe

relationships between environmental variables and

species distribution and/or abundance. This is done

using a wide array of statistical methods that relate

field observations to environmental predictors.

Nowadays, digital land use maps are available over

large areas. Such maps classify the land cover

according to a large number of classes, which are

frequently used as environmental predictors, often

with topographic and/or climatic variables (e.g.

Seoane et al. 2003; Posillico et al. 2004; Virkkala et al.

2005; Tapia et al. 2007). Nevertheless, HSMs based

on land use, topographic and climatic predictors may

sometimes show large difference between predicted

and observed values (Massimino 2004). This can be

due to inadequacy of the modelling technique,

unsaturated habitats (Cardillo et al. 1999; Tyre et al.

2001) or failure to include significant environ-

mental predictors (Crawley 1993; Guisan & Thuiller

2005), whose selection is often critical and strongly

dependent on the type of selection strategy used

(Araujo & Guisan 2006). Moreover, occurrence of

many species can be influenced by environmental

characteristics or processes that are difficult to

measure, or at least have not been measured

adequately. In the case of forest species, the environ-

mental characteristics influencing their presence and

abundance can be the amount of forest cover, size

and isolation of forest patches, edge effect, and age

and structure of stands (Villard et al. 1999; Hobson &

Bayne 2000; Giese & Cuthbert 2003; Kurosawa &

Askins 2003; Batari & Baldi 2004; Herrmann et al.

2005; Venier & Pearce 2005; Bani et al. 2006). While

forest cover and landscape and patch metrics can be

obtained or derived from land-use digital maps, data

on forest structure are rarely available over large areas.

*Correspondence: L. Bani, Department of Environmental and Landscape Sciences, University of Milano-Bicocca, Piazza della Scienza 1, Milano 20126, Italy.

Tel: +39 02 64482944. Email: luciano.bani@unimib.it

Italian Journal of Zoology, March 2010; 77(1): 71–80

ISSN 1125-0003 print/ISSN 1748-5851 online # 2010 Unione Zoologica Italiana

DOI: 10.1080/11250000802658058

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

In the case of Lombardy (northern Italy), data on

forest structure are not available over the whole

region. The need to produce reliable habitat

suitability models for bird species prompted us to

test the use of coarse grain data (available only at

province level) on the percentage of forest manage-

ment systems, which are the major determinant of

habitat structure in our forests, that are all subjected

to silvicultural practices. We also used an index of

forest patch isolation (Bani et al. 2006) derived from

the proximity index calculated by the FRAGSTATS

program (McGarigal & Marks 1995).

The species considered were the long-tailed tit

Aegithalos caudatus (Linnaeus, 1758), marsh tit

Poecile palustris (Linnaeus, 1758), blue tit Cyanistes

caeruleus (Linnaeus, 1758), wood nuthatch Sitta

europaea Linnaeus, 1758 and short-toed tree-creeper

Certhia brachydactyla Brehm, 1820. They all are

typical of broadleaf forests and sensitive to fragmen-

tation, but show different selectivity to the forest

structure (Bani et al. 2006). In our study area the

marsh tit, wood nuthatch and short-toed tree-

creeper need mature forests, the blue tit can also

use intermediate seral stages, while the long-tailed tit

may breed in forests of any age (Bani et al. 2006).

Materials and methods

Study area

Our study was carried out in the Lombardy region

(northern Italy). Approximately 25% of the study

area is covered by forests, of which 63% are

broadleaf, 22% coniferous and 15% mixed. The

great majority (94%) of broadleaf forests, the focus

of this research, are managed as coppice, but there

are considerable differences among provinces in the

proportion of compound coppice (coppice with

standards of several age classes): in the western

provinces compound coppice prevails, while in the

central and eastern provinces simple coppice (with

or without standards) is predominant (Figure 1).

Faunal and environmental data

Bird data were collected by means of the unlimited

distance point-count technique, which consists of

recording each bird seen or heard in a 10-min

session (Blondel et al. 1981). This technique

provides values of relative bird abundance (Blondel

et al. 1970; Bibby et al. 2000). Point counts were

conducted during the breeding season from sunrise

to 11 am in good weather (sunny to cloudy, without

rain or strong wind) (Fornasari et al. 1998). From

1992 to 2002 we performed 4552 sample points

evenly distributed across the entire study area.

Distances between sampling locations were at least

1000 m.

Land-use variables were derived from the

1:10,000 DUSAF (Destinazione d’Uso dei Suoli

Agricoli e Forestali [Classification of Agricultural

and Forest Lands]; ERSAF 2002) digital map with

20-m ground resolution. They were measured as the

fractional cover in a 500-m radius circle centred on

each sample point. Since the classification of non-

forest areas was too detailed for our aims, we

aggregated the original 61 land-use classes into 16

main classes. We added an interaction term between

all forests and all open environments to account for

changes in suitability at ecotones.

Topographic variables were obtained from a

Digital Elevation Model. They were elevation,

square elevation, slope and exposure, the last

calculated as the sine and cosine of the aspect.

Forest patch isolation was also derived from the

DUSAF digital map. It was measured by the

FRAGSTATS Proximity Index (McGarigal &

Marks 1995), modified by Bani et al. (2006). The

FRAGSTATS Proximity Index of a patch equals

zero if there are no other patches of the same

category (in our case broadleaf forest patches) within

the proximity radius and increases as the area

occupied by surrounding patches increases and/or

their distances from the focal patch decrease. The

Modified Proximity Index (MPI; Bani et al. 2006)

for a patch is defined as the product of its Proximity

Index and area. In our case study we used a 2500-m

proximity radius, which avoided considering neigh-

bouring forest patches separated by the highest

mountain ridges, a potentially important barrier

during animal dispersal. The values of the natural

logarithm of the MPI were calculated for each

broadleaf forest patch in the study area and then

assigned to sample points contained in them. Points

within 500 m from a broad-leaved forest patch were

assigned its proximity value, while points more than

500 m far from a broadleaf forest were assigned a

value of zero.

Data on forest management systems (see

Figure 1), updated in 1997 and only available at

provincial level for each of the former 11 provinces

of Lombardy, came from the Regional Statistical

Yearbook (1997). We assumed that in each point the

apportionment of broad-leaved forests in each

management system were the same throughout the

province. Therefore we assigned to each sample

point the fraction of broadleaf forests managed,

respectively, as simple coppice, compound coppice

and high forest, according to the province where the

point was located. We then multiplied the fractions

72 D. Massimino et al.

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

of management types by the broadleaf forest frac-

tional cover in the 500-m radius circle around each

point. In this way, we split the broadleaf forest land-

use variable in three different variables according to

the different management types.

Table I summarizes the environmental predictor

variables used, their abbreviations and ranges in the

study area.

Statistical analysis

To model the habitat suitability for each considered

species, we used generalized linear models (GLMs),

which have proven effective in several studies (e.g.

Martinez et al. 2003; Thuiller 2003; Randin et al.

2006; Brotons et al. 2007; Rodriguez-Estrella 2007).

Models were performed with negative binomial error

distribution and log link function, as are appropriate

for overdispersed count data (Bliss & Fisher 1953;

Quinn & Keough 2002):

E yð Þ~exp Sbxð Þ

where E(y) is the expected value of the dependent

variable (number of birds counted in each point

count) and S bx is the linear predictor, a linear

Figure 1. Broadleaf forest surface in each Lombardy province. The bar height is proportional to the surface and numbers below the bars are

the percentages of each management system.

Forest management and bird abundance 73

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

combination of parameters b and independent

variables x, which are the environmental variables

listed in Table I.

To obtain a good and parsimonious model,

independent variables were selected by a stepwise

procedure based on the Bayesian Informative

Criterion (BIC), which chose the model that

minimizes the log-likelihood minus the number of

parameters multiplied by the logarithm of the sample

size (see Burnham & Anderson 2002). Generalized

Linear Regression and BIC stepwise selection of

variables were performed using S-PLUS software

Version 7.0 (S-PLUS 2005).

Model goodness-of-fit was evaluated by residual

deviance, which expresses the discrepancy between

observed and fitted values (see Quinn & Keough

2002). The lower the ratio between the residual and

null deviance, the better is the model capacity to fit

the data.

For each of the five species, we cross-validated the

GLMs by a leave-more-out procedure (see Guisan &

Thuiller 2005), randomly splitting the samples into

four equal-sized subsets. Each GLM was then run

four times, using three subsets as the training set and

the other as the test set. We computed the residual

deviance as the discrepancy between the fitted values

estimated by the model built on the training set and

the observed values in the test sets. The residual

deviance so calculated was then compared to the

residual deviance of the suitability GLM to assess

the predictive ability of the model (Vernier et al.

2002).

Equations resulting from GLMs were used to

draw habitat suitability maps by using a

Geographical Information System.

Results

The stepwise procedure based on the BIC selected

final models with 7–14 independent variables,

explaining between 17.9 and 43.0% of the null

deviance. Similar values were obtained in validation,

with a loss of deviance explained ranging between

1.2 and 5.0%. Model summaries (null deviance,

residual deviance, percent of deviance explained by

the model residual deviance in validation and

percent of deviance explained in validation) are

given in Table II.

Table III shows the environmental predictor vari-

ables for each species and their coefficients, standard

errors and t values. The marsh tit, blue tit, wood

nuthatch and short-toed tree-creeper had positive

coefficients for the MPI, broadleaf forests managed

as high forest and either a positive coefficient for

broadleaf forests managed as compound coppice or

a negative coefficient for those managed as simple

coppice. The long-tailed tit did not exhibit any

preference with respect either to forest management

or patch isolation.

Table I. Environmental predictor variables and their range in the study area.

Abbreviation Description Range in the study area

E Elevation 0.005–3.996 km

C Cosine of the aspect 21–1

S Sine of the aspect 21–1

P Slope 0–84.6uART Artificial surfaces 0–1

ARA Non-irrigated arable land 0–1

RIC Rice fields 0–1

PER Permanent crops 0–1

PAS Pastures 0–1

HET Heterogeneous agricultural areas 0–1

BROSEMP Broad-leaved forests managed as simple coppice 0–0.91

BROCOMP Broad-leaved forests managed as compound coppice 0–0.76

BROHIGH Broad-leaved forests managed as high forest 0–0.23

CON Coniferous forests 0–1

MIX Mixed forests 0–1

GRA Natural grasslands 0–1

MOO Moors and heathland 0–1

WAT Water courses, beaches, dunes, sands 0–1

ROC Bare rock 0–1

SPA Sparsely vegetated areas 0–1

WET Wetlands 0–1

O All open areas landscapes 0–1

F All forests 0–1

Ln(MPI) Natural logarithm of the Modified Proximity Index 0–17.0

74 D. Massimino et al.

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

The maps of habitat suitability and abundance of

detected birds for each species are given in Figure 2.

Discussion

In this research we tested the use of coarse grain data

on forest management systems, which are major

determinants of forest structure in Lombardy, com-

bined with land use, topographic and landscape

variables to model the habitat suitability of five forest

species. Although data on forest management were

available only as overall values over each of the former

11 provinces, they were effective in improving

performance of habitat suitability models for four

out of five species. We can infer the importance of

forestry data from the inclusion in the final model of

variables concerning forest management. Their coef-

ficients, estimated by the GLM, were significantly

different, which means that broad-leaved forests have

different suitability according to their management.

Since bird data were collected during an 11-year

period, our models rely on the assumption that no

important changes occurred in land use and forest

management. Although detailed data are not avail-

able year-by-year, percent changes in urban, agri-

cultural and forest areas in our study area were less

than 3% between 1992 and 2005 (ARPA 2002,

2006) and changes in forest management systems

were around 0.06% between 1992 and 1997, the last

year for which forest data are available (Regional

Statistical Yearbook 1997).

Of all common forest species breeding in the study

area, the five considered here were the most sensitive

to anthropogenic disturbances to forests (Bani et al.

2006). Among them, marsh tit, wood nuthatch and

short-toed tree-creeper, which are cavity nesters

(Harrap & Quinn 1996), were proven to be the

most selective for forest structure (Bani et al. 2006).

For these three species, the stepwise procedure used

in this research selected both the variable of broad-

leaf forests managed as high forest and the variable

of broadleaf forests managed as compound coppice.

High forests had quite obviously the highest coeffi-

cient. Nevertheless, the maximum fractional cover of

this variable throughout the region is 0.23. On the

other hand, compound coppices, which are fairly

common in the western part of the study area, had

lower but highly significant (P , 0.0001 in all three

models) positive coefficients. It results in remarkable

differences between abundances estimated in com-

pound coppice and in simple coppice forests. All

other variables being equal, in areas completely

covered with broadleaf forests the abundance of

short-toed tree-creeper, marsh tit and wood

nuthatch is estimated, respectively, 69.9, 21.7 and

9.1 times greater in the case of compound coppice

than in simple coppice.

This confirms that data on forest management

systems, although aggregated over provinces, are

useful to improve habitat suitability models for these

structure-sensitive species, whose abundance is

strongly affected by silvicultural practices.

The blue tit is also a cavity nester, therefore rather

selective for the forest structure. However, when

natural tree cavities are scarce, it is able to use other

natural or artificial sites, such as wall crevices or

ground holes (Harrap & Quinn 1996). The stepwise

procedure selected the simple coppice variable with

a negative coefficient. Its abundance in areas

completely covered with broadleaf forest is estimated

2.6 times greater in compound coppice than in

simple coppice.

These four species were also affected by forest

fragmentation. Their environmental suitability mod-

els included the logarithm of MPI with a significant

positive coefficient.

In the case of the long-tailed tit model only, no

variable concerning forest management was selected.

Although in other areas the long-tailed tit was

proven to be selective for forest structure, proportion

and isolation of suitable habitat (Gaston 1973;

Jansson & Angelstam 1999), in our study area it

seemed to be rather generalist. The species did not

appear to be affected even by fragmentation, as the

logarithm of MPI was excluded by the stepwise

procedure. This may suggest a higher dispersal

ability of the long-tailed tit compared to the other

four species (Harrap & Quinn 1996).

The results obtained in this research have practical

implications for the conservation of forest species.

Table II. Summaries of the habitat suitability models for each species.

Species Null deviance Residual deviance

Deviance explained by

the model (%)

Residual deviance in

validation

Deviance explained in

validation (%)

Long-tailed tit 2200.0 1806.5 17.9 1873.4 14.8

Marsh tit 2548.8 1511.1 40.7 1569.1 38.4

Blue tit 5074.6 3500.3 31.0 3560.6 29.8

Wood nuthatch 2146.6 1318.0 38.6 1395.7 35.0

Short-toed tree-creeper 1368.1 779.3 43.0 848.3 38.0

Forest management and bird abundance 75

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

Table III. Environmental predictor variables selected by the stepBIC procedure for each species and their coefficients, standard errors and t-values.

Species Variable Coefficient SE t-value

Long-tailed tit E 2.917 0.496 5.880

E2 21.948 0.247 27.880

ARA 21.766 0.289 26.121

RIC 21.909 0.372 25.130

HET 22.051 0.384 25.343

GRA 23.162 0.569 25.562

O * F 6.169 0.832 7.413

(Intercept) 21.748 0.211 28.273

Marsh tit ARA 21.552 0.422 23.679

RIC 2425.087 357.147 21.190

HET 23.011 0.941 23.200

BROCOMP 3.079 0.417 7.384

BROHIGH 11.281 1.624 6.945

MIX 2.523 0.293 8.624

GRA 24.485 0.702 26.388

MOO 24.569 1.167 23.914

WAT 27.747 1.958 23.957

MPI 0.063 0.015 4.204

O * F 7.620 1.051 7.247

(Intercept) 24.235 0.223 218.975

Blue tit E 3.651 0.452 8.080

E2 22.391 0.250 29.545

P 20.026 0.006 24.137

ART 21.547 0.272 25.687

ARA 22.360 0.228 210.341

RIC 21.894 0.277 26.837

HET 23.406 0.411 28.292

BROSEMP 20.952 0.250 23.808

BROHIGH 6.975 1.138 6.123

CON 21.163 0.298 23.896

GRA 21.784 0.349 25.112

MOO 22.183 0.658 23.317

MPI 0.037 0.009 4.129

O * F 3.194 0.579 5.521

(Intercept) 21.757 0.170 210.330

Wood nuthatch E 20.736 0.240 23.067

S 0.534 0.142 3.754

P 20.032 0.009 23.544

ARA 23.136 0.571 25.493

RIC 2453.841 348.147 21.304

HET 25.320 1.402 23.793

BROCOMP 2.205 0.441 4.996

BROHIGH 11.999 1.654 7.253

MIX 2.057 0.314 6.556

GRA 22.311 0.761 23.037

MOO 24.446 1.483 22.998

WAT 27.855 1.900 24.133

MPI 0.082 0.018 4.669

O * F 7.014 1.176 5.962

(Intercept) 23.664 0.270 213.578

Short-toed tree-creeper P 20.101 0.010 210.038

ARA 22.100 0.656 23.200

RIC 2295.089 234.539 21.258

HET 24.343 1.618 22.684

BROCOMP 4.248 0.567 7.485

BROHIGH 10.118 2.260 4.476

MIX 2.968 0.423 7.010

GRA 24.055 1.117 23.630

WAT 218.509 4.447 24.162

MPI 0.104 0.024 4.378

O * F 8.14 1.378 5.909

(Intercept) 24.687 0.369 212.712

76 D. Massimino et al.

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

Although the overall area of forest in Europe has

greatly increased since the 1960s (EEA 2003),

ecologically high-quality old-growth forests have

been almost completely replaced by intensively

managed forests or plantations whose structure is

poorly suited to host most cavity-nesting species

Figure 2. Maps of habitat suitability and abundance of birds detected in each point count. Dashed areas are major lakes.

Forest management and bird abundance 77

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

(Tucker & Evans 1997), which are likely to suffer

detrimental impacts and population declines.

Among the species considered in this research, the

marsh tit underwent the largest decline in north-west

Europe during 1990–2000, so that it was included

among the Species of European Conservation

Concern (BirdLife International 2004). Wood

nuthatch and short-toed tree-creeper seem more

stable, although the former’s sizeable population in

France suffered a large decline (255%) in 1990–

2000 (BirdLife International 2004). To combat their

decline, semi-natural forests should be restored to

natural old-growth conditions where possible

(Tucker & Evans 1997). Unfortunately this is

unfeasible in the intensively exploited forests of the

study area, but a pragmatic objective could be

restoring natural features such as mixtures of

different tree age classes, some old-growth trees

and dead wood. At a minimum, simple coppices

should be converted into compound coppices, which

appear more suitable for the most selective cavity

nesters.

Acknowledgements

We thank the Agricultural and Forest Service of

Lombardy Regional Administration for financing

bird surveys. We are also grateful to three anon-

ymous referees for their helpful suggestions.

References

Araujo MB, Cabeza M, Thuiller W, Hannah L, Williams PH.

2004. Would climate change drive species out of reserves? An

assessment of existing reserve-selection methods. Global

Change Biology 10:1618–1626.

Araujo MB, Guisan A. 2006. Five (or so) challenges for species

distribution modelling. Journal of Biogeography 33:1677–

1688.

Figure 2. Continued.

78 D. Massimino et al.

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

ARPA. 2002. Segnali ambientali della Lombardia. Rapporto sullo

stato dell’ambiente-2002. Milano: Agenzia Regionale per la

Protezione dell’Ambiente della Lombardia, Regione

Lombardia.

ARPA. 2006. Rapporto sullo stato dell’ambiente. Edizione 2006.

Milano: Agenzia Regionale per la Protezione dell’Ambiente

della Lombardia, Regione Lombardia.

Bani L, Massimino D, Bottoni L, Massa R. 2006. A multiscale

method for selecting indicator species and priority conserva-

tion areas: A case study for broadleaved forests in Lombardy,

Italy. Conservation Biology 20:512–526.

Batari P, Baldi A. 2004. Evidence of an edge effect on avian nest

success. Conservation Biology 18:389–400.

Berry PM, Dawson TP, Harrison PA, Pearson RG. 2002.

Modelling potential impacts of climate change on the

bioclimatic envelope of species in Britain and Ireland. Global

Ecology and Biogeography 11:453–462.

Bibby CJ, Burgess ND, Hill DA, Mustoe S. 2000. Bird census

techniques. London: Academic Press.

BirdLife International. 2004. Birds in Europe: Population

estimates, trends and conservation status. Cambridge:

BirdLife International. 374 p.

Bliss CI, Fisher RA. 1953. Fitting the negative binomial

distribution to biological data. Biometrics 9:176–200.

Blondel J, Ferry C, Frochot B. 1970. La methode des indices

ponctuels d’abondance (I.P.A.) ou des releves d’avifaune par

‘‘stations d’ecoute’’. Alauda 38:55–71.

Blondel J, Ferry C, Frochot B. 1981. Point counts with unlimited

distance. Studies in Avian Biology 6:414–420.

Brotons L, Herrando S, Pla M. 2007. Updating bird species

distribution at large spatial scales: Applications of habitat

modelling to data from long-term monitoring programs.

Diversity and Distributions 13:276–288.

Burnham KP, Anderson DR. 2002. Model selection and multi-

model inference: A practical information-theoretic approach.

2nd ed. New York: Springer-Verlag.

Cardillo M, Macdonald DW, Rushton SP. 1999. Predicting

mammal species richness and distributions: Testing the

effectiveness of satellite-derived land cover data. Landscape

Ecology 14:423–435.

Crawley MJ. 1993. GLIM for ecologists. Oxford: Blackwell Science.

EEA. 2003. Europe’s environment. The third assessment.

Environmental assessment report no. 10. Copenhagen:

European Environment Agency. 344 p.

ERSAF. 2002. Destinazione d’Uso dei Suoli Agricoli e Forestali

(DUSAF). Milano: Ente Regionale per i Servizi all’Agricoltura

e alle Foreste della Lombardia.

Fornasari L, Bani L, de Carli C, Massa R. 1998. Optimum design

in monitoring common birds and their habitat. Gibier Faune

Sauvage 15:309–322.

Gaston AJ. 1973. The ecology and behaviour of the Long-tailed

Tit. Ibis 115:330–351.

Giese CLA, Cuthbert FJ. 2003. Influence of surrounding

vegetation on woodpecker nest tree selection in oak forests of

the Upper Midwest, USA. Forest Ecology and Management

179:523–534.

Guisan A, Thuiller W. 2005. Predicting species distribution:

Offering more than simple habitat models. Ecology Letters

8:993–1009.

Harrap S, Quinn D. 1996. Tits, nuthatches and treecreepers.

London: Helm.

Herrmann HL, Babbitt KJ, Baber MJ, Congalton RG. 2005.

Effects of landscape characteristics on amphibian distribution

in a forest-dominated landscape. Biological Conservation

123:139–149.

Hobson KA, Bayne E. 2000. The effects of stand age on avian

communities in aspen-dominated forests of central

Saskatchewan, Canada. Forest Ecology and Management

136:121–134.

Jansson G, Angelstam P. 1999. Threshold levels of habitat

composition for the presence of the long-tailed tit (Aegithalos

caudatus) in a boreal landscape. Landscape Ecology

14:283–290.

Kurosawa R, Askins RA. 2003. Effects of habitat fragmentation

on birds in deciduous forests in Japan. Conservation Biology

17:695–707.

Maiorano L, Falcucci A, Boitani L. 2006. Gap analysis of

terrestrial vertebrates in Italy: Priorities for conservation

planning in a human dominated landscape. Biological

Conservation 133:455–473.

Martinez JA, Serrano D, Zuberogoitia I. 2003. Predictive models

of habitat preferences for the Eurasian eagle owl Bubo bubo: A

multiscale approach. Ecography 26:21–28.

Massimino D. 2004. Valutazione della qualita dei corpi forestali

dell’area orobica mediante bioindicatori [dissertation].

Milano: University of Milano-Bicocca, 129 p. Available:

Biblioteca Nazionale Centrale di Firenze; CF020041625.

McGarigal K, Marks BJ. 1995. FRAGSTATS: Spatial pattern

analysis program for quantifying landscape structure. General

technical report PNW-GTR-351. Portland: U.S. Department

of Agriculture Forest Service, Pacific Northwest Research

Station.

Posillico M, Alberto MIB, Pagnin E, Lovari S, Russo L. 2004. A

habitat model for brown bear conservation and land use

planning in the central Apennines. Biological Conservation

118:141–150.

Puglisi L, Adamo MC, Baldaccini NE. 2005. Man-induced

habitat changes and sensitive species: A GIS approach to the

Eurasian Bittern (Botaurus stellaris) distribution in a

Mediterranean wetland. Biodiversity and Conservation

14:1909–1922.

Quinn JP, Keough MJ. 2002. Experimental design and data

analysis for biologists. Cambridge: Cambridge University

Press.

Randin CF, Dirnbock T, Dullinger S, Zimmermann NE,

Zappa M, Guisan A. 2006. Are niche-based species distribu-

tion models transferable in space? Journal of Biogeography

33:1689–1703.

Regional Statistical Yearbook. 1997. Agriculture, forestry, hunt-

ing and fisheries. Forest areas and wood production. Milano:

Regione Lombardia, Unioncamere Lombardia, ISTAT.

Available: http://www.ring.lombardia.it/asrnewe/index.html.

Accessed Feb 2008 22.

Rodriguez-Estrella R. 2007. Land use changes affect distribu-

tional patterns of desert birds in the Baja California peninsula,

Mexico. Diversity and Distributions 13:877–889.

Rondinini C, Chiozza F, Boitani L. 2006. High human density in

the irreplaceable sites for African vertebrates conservation.

Biological Conservation 133:358–363.

Seoane J, Justribo JH, Garcia F, et al. 2006. Habitat-suitability

modelling to assess the effects of land-use changes on

Dupont’s lark Chersophilus duponti: A case study in the Layna

Important Bird Area. Biological Conservation 128:241–252.

Seoane J, Vinuela J, Diaz-Delgado R, Bustamante J. 2003. The

effects of land use and climate on red kite distribution in the

Iberian peninsula. Biological Conservation 111:401–414.

S-PLUS. 2005. S-PLUS 7.0 for Windows. Seattle: Insightful

Corporation.

Tapia L, Dominguez J, Rodriguez L. 2007. Modelling habitat

use and distribution of golden eagles Aquila chrysaetos in a

Forest management and bird abundance 79

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010

low-density area of the Iberian Peninsula. Biodiversity and

Conservation 16:3559–3574.

Thuiller W. 2003. BIOMOD – optimizing predictions of species

distributions and projecting potential future shifts under global

change. Global Change Biology 9:1353–1362.

Tucker GM, Evans MI. 1997. Habitats for birds in Europe: A

conservation strategy for the wider environment. Cambridge:

BirdLife International.

Tyre AJ, Possingham HP, Lindenmayer DB. 2001. Inferring

process from pattern: can territory occupancy provide infor-

mation about life history parameters? Ecological Applications

11:1722–1737.

Venier LA, Pearce JL. 2005. Boreal bird community response to jack

pine forest succession. Forest Ecology and Management 217:19–36.

Vernier PR, Schmiegelow FKA, Cumming SG. 2002. Modeling

bird abundance from forest inventory data in the boreal

mixed-wood forests of Canada. In: Scott JM, Heglund PJ,

Morrison ML, Haufler JB, Raphael MG, Wall WA,

Samson FB, editors. Predicting species occurrences. Issues

of accuracy and scale. Washington: Island Press. pp. 559–

571.

Villard MA, Trzcinski MK, Merriam G. 1999. Fragmentation

effects on forest birds: Relative influence of woodland cover

and configuration on landscape occupancy. Conservation

Biology 13:774–783.

Virkkala R, Luoto M, Heikkinen RK, Leikola N. 2005.

Distribution patterns of boreal marshland birds: Modelling

the relationships to land cover and climate. Journal of

Biogeography 32:1957–1970.

Zielinski WJ, Carroll C, Dunk JR. 2006. Using landscape

suitability models to reconcile conservation planning for two

key forest predators. Biological Conservation 133:409–430.

80 D. Massimino et al.

Downloaded By: [Bani, Luciano] At: 12:57 12 March 2010