Ss

HPCIa

b

c

d

2e

f

4g

h

i

j

a

AA

KBECP

pv

h0

International Journal of Applied Earth Observation and Geoinformation 37 (2015) 124–132

Contents lists available at ScienceDirect

International Journal of Applied Earth Observation andGeoinformation

jo ur nal home p age: www.elsev ier .com/ locate / jag

atellite Earth observation data to identify anthropogenic pressures inelected protected areas

arini Nagendraa,∗, Paola Mairotab, Carmela Marangic, Richard Lucasd,anayotis Dimopoulose, João Pradinho Honradof, Madhura Niphadkara,aspar A. Mücherg, Valeria Tomasellih, Maria Panitsae, Cristina Tarantino j,

oannis Manakos i, Palma Blondaj

Ashoka Trust for Research in Ecology and the Environment (ATREE), Royal Enclave, Srirampura, Jakkur Post, Bangalore 560064, Karnataka, IndiaDipartimento di Scienze Agro-Ambientali e Territoriali, Università degli Studi di Bari “Aldo Moro”, Via Orabona, 4–70126 Bari, ItalyNational Research Council- Istituto per le Applicazioni del Calcolo, Via Amendola 122/D, 70126, Bari, ItalyCentre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, the University of New South Wales, High Street, Kensington, NSW052, AustraliaDepartment of Environmental and Natural Resources Management, University of Patras, Seferi 2, 30100 Agrinio, GreeceCIBIO-Centro de Investigac ão em Biodiversidade e Recursos Genéticos & Faculdade de Ciências, Universidade do Porto, Rua do Campo Alegre, s/n,169-007 Porto, PortugalAlterra, Wageningen UR, Droevendaalsesteeg 3, 6708 PB, Wageningen, The NetherlandsNational Research Council – Institute of Biosciences and BioResources (CNR-IBBR), Via G. Amendola 165/A, 70126 Bari, ItalyInformation Technologies Institute (ITI), Centre for Research & Technology Hellas, 6th km Harilaou – Thermi, 57001 Thessaloniki, GreeceNational Research Council – Institute of Intelligent Systems for Automation (CNR-ISSIA), Via G. Amendola 122, 70126 Bari, Italy

r t i c l e i n f o

rticle history:vailable online 22 November 2014

eywords:iodiversity conservationarth observationhanges in staterotected areas

a b s t r a c t

Protected areas are experiencing increased levels of human pressure. To enable appropriate conserva-tion action, it is critical to map and monitor changes in the type and extent of land cover/use and habitatclasses, which can be related to human pressures over time. Satellite Earth observation (EO) data andtechniques offer the opportunity to detect such changes. Yet association with field information and expertinterpretation by ecologists is required to interpret, qualify and link these changes to human pressure.There is thus an urgent need to harmonize the technical background of experts in the field of EO dataanalysis with the terminology of ecologists, protected area management authorities and policy makers inorder to provide meaningful, context-specific value-added EO products. This paper builds on the DPSIRframework, providing a terminology to relate the concepts of state, pressures, and drivers with the appli-cation of EO analysis. The type of pressure can be inferred through the detection of changes in state (i.e.changes in land cover and/or habitat type and/or condition). Four broad categories of changes in stateare identified, i.e. land cover/habitat conversion, land cover/habitat modification, habitat fragmentationand changes in landscape connectivity, and changes in plant community structure. These categories ofchange in state can be mapped through EO analyses, with the goal of using expert judgement to relatechanges in state to causal direct anthropogenic pressures. Drawing on expert knowledge, a set of pro-tected areas located in diverse socio-ecological contexts and subject to a variety of pressures are analysedto (a) link the four categories of changes in state of land cover/habitats to the drivers (anthropogenic pres-sure), as relevant to specific target land cover and habitat classes; (b) identify (for pressure mapping) themost appropriate spatial and temporal EO data sources as well as interpretations from ecologists and

field data useful in connection with EO data analysis. We provide detailed examples for two protectedareas, demonstrating the use of EO data for detection of land cover/habitat change, coupled with expertinterpretation to relate such change to specific anthropogenic pressures. We conclude with a discussionof the limitations and feasibilisuggesting additional research

∗ Corresponding author. Tel.: +91 80 23635555; fax: +91 80 23530070.E-mail addresses: nagendra@atree.org (H. Nagendra), paola.mairota@uniba.it (P. Ma

dimopoulos@upatras.gr (P. Dimopoulos), jhonrado@fc.up.pt (J.P. Honrado), madhualeria.tomaselli@ibbr.cnr.it (V. Tomaselli), mpanitsa@upatras.gr (M. Panitsa), imanakos@

ttp://dx.doi.org/10.1016/j.jag.2014.10.010303-2434/© 2014 Elsevier B.V. All rights reserved.

ty of using EO data and techniques to identify anthropogenic pressures, efforts required in this direction.

© 2014 Elsevier B.V. All rights reserved.

irota), c.marangi@ba.iac.cnr.it (C. Marangi), richard.lucas@unsw.edu.au (R. Lucas),ra.niphadkar@atree.org (M. Niphadkar), sander.mucher@wur.nl (C.A. Mücher),iti.gr (I. Manakos), blonda@ba.issia.cnr.it (P. Blonda).

arth O

I

aiwtahst

it2sEAavt

apsotamatddoe

fihaEpdtapia

Uv(mb

M

Ibiadeit

H. Nagendra et al. / International Journal of Applied E

ntroduction

Local scale anthropogenic pressures on the Earth have led toccelerated declines in biodiversity (Hooper et al., 2012), affect-ng the provisioning of ecosystem services essential for human

ellbeing (Mace et al., 2012). Thus, international bodies such ashe recently established Intergovernmental Panel on Biodiversitynd Ecosystem Services (IPBES) have stressed the need to assessuman pressures on biodiversity and ecosystem services across allcales, including at local scales relevant to management interven-ion (Honrado et al., 2013).

Although protected areas constitute a cornerstone of local andnternational conservation approaches (Jenkins and Joppa, 2009),hey continue to experience anthropogenic pressure (Nagendra,008; Vicente et al., 2013). For effective management response,patial knowledge of the type and location of pressure is required.O and airborne data from passive optical and active Syntheticperture Radar (SAR) and Light Detection and Ranging (LiDAR) arevailable at different spectral and spatial resolutions, and can beery useful for detecting land cover/use and/or habitat changes overime (Nagendra et al., 2013).

EO and associated techniques, coupled with landscape patternnalysis and habitat modelling of biodiversity distributions, canrovide information on changes in the state/condition of biodiver-ity and ecosystems. These changes can be used to infer evidencef pressures. However, there is an urgent need to harmonize theechnical terminology of experts in the field of EO data analysisnd of ecologists, with the applied scientific terminologies of policyakers and management authorities. Policy makers and protected

rea management authorities use terms such as impact, pressures,hreats, drivers, state, stress–response (Salafsky et al., 2008). It isifficult to directly relate these terms to corresponding indicatorserived from EO data analysis, which mainly provides informationn the extent and location of changes in land cover/use (Tarantinot al., 2007).

The aim of this paper is to harmonize the discourse betweeneld experts, ecologists and protected area managers on the oneand, and EO experts on the other hand, by proposing a unifiedpproach to facilitate the provision of value-added products fromO sources that can be useful for biodiversity conservation pur-oses. This is illustrated for selected protected areas, located iniverse socio-ecological contexts and representing diverse habitatypes including wetlands, forests, grasslands and bogs subject to

variety of pressures, with two detailed examples provided. Theaper concludes with a brief discussion of the advantages and lim-

tations of EO data techniques for pressure assessment in protectedreas.

This study was developed and tested within the Europeannion’s Seventh Framework Programme (EU-FP7) project Biodi-ersity Multi-SOurce Monitoring System: From Space To SpeciesBIO SOS), that aimed to develop tools and models for consistent

ulti-annual monitoring of protected areas and their surroundingsy the integrated use of EO and in-field data.

ethods: definition and categorization of changes in state

We propose to build on the Driving forces, Pressure, State,mpact and Response (DPSIR) framework (EEA, 1995), defining fourroad categories of changes in state which can be mapped and mon-

tored through EO analyses. Changes in state are likely to be site-nd habitat-specific (Nagendra et al., 2012). These are described,

rawing on a systematic analysis based on literature review andxpert knowledge (judgement), for a number of Natura 2000 sitesn Europe and a protected area in southern India. The sites are fur-her described in Table 1. European sites are located in countries

bservation and Geoinformation 37 (2015) 124–132 125

belonging to the Mediterranean (Italy, Greece and Portugal) andAtlantic (Portugal, The Netherlands and Wales) biogeographicalregions (Habitats Directive–92/43/EEC). For each site, interpreta-tion from ecologists, drawing on field data, is used to link specificchanges in state with the direct pressures considered as drivers inthe protected area. Appropriate EO data are then used to analysespecific changes in state. Detailed examples are provided for twoprotected area sites, illustrating the use of EO data for detectingchanges in state related to pressures.

Taxonomy for pressure assessment

A number of different terms have been used by researchers andconservation agencies to define pressures and consequent changesin state of habitats in protected area assessments. Perhaps themost widely used by policy makers and managers is the DPSIRframework, although there are a number of other frameworksand approaches such as strengths, weaknesses, opportunities andthreats (SWOT) analysis (widely used in other contexts), theWCPA/IUCN framework for assessment and monitoring of pro-tected areas (Hockings et al., 2006), and the work of Salafsky et al.(2008), reviewed further in Nagendra et al. (2012).

According to the DPSIR framework, Drivers of environmentalchange (e.g., urbanization) result in Pressures on the environment(e.g., discharge of industrial and domestic waste), which createchanges in the State of the environment (e.g., water quality), cre-ating Impacts (e.g., modifications of ecological functions and inecosystem services provision) on ecology and society. This leadsto Responses (e.g., policy/management responses) – which in turnfeedback on Drivers, thus reinstating the overall cycle. Thus theDPSIR framework, through state and impact monitoring, allowsfor the assessment of response effectiveness and appropriatenesswithin an adaptive management perspective.

Pressures in the DPSIR framework refer to processes of disturb-ance that influence changes in biodiversity and ecosystem state.It is difficult to directly detect pressures, or to assess impacts ofsuch pressures on society and ecology through EO. EO can howeverdetermine changes in land cover/use or habitat e.g. in forest struc-ture (state of the forest) as a consequence of pressure (Nagendraand Rocchini, 2008). Through expert knowledge, changes in statecan be used to infer both the type of disturbance (e.g., wildfire, windstorm, pest infestation, silviculture, overharvesting) as a drivingpressure, and the type of impacts likely to ensue, so that appropriateresponses can be devised (e.g. Nagendra et al., 2010).

Broad categorization of changes in state

Based on literature review and the knowledge of ecologists,four broad categories of changes in state are identified, which canbe mapped through EO data analyses, and related to causal directanthropogenic pressures based on expert ecological interpretation.These are

1 Land cover/habitat conversion;2 Land cover/habitat modification (e.g., in vegetation structure, as

indicated by changes in density or cover, height);3 Habitat fragmentation and changes in landscape connectivity;4 Changes in plant community structure.

In some instances land cover/use classes (e.g. deciduous forest)coincide with habitat categories, but in general the definition of ahabitat (sensu Bunce et al., 2013) is linked to that of an ecosystem,

and includes additional ecological concepts and components (e.g.,lithology, slope, water salinity).

Expert knowledge is crucial in the EO analytical phase. Ecologistsand other field experts rely on processes of human discrimination

126

H.

Nagendra

et al.

/ International

Journal of

Applied

Earth O

bservation and

Geoinform

ation 37

(2015) 124–132

Table 1Description of study sites.

Natura 2000 site name Natura2000 Code

Habitats of conservation concern (*) General description

Rios Sabor and Mac ãs PTZPE0037andPTCON0021

Intermittently flowing Mediterranean rivers of the Paspalo-Agrostidion(3290)Stable xerothermophilous formations with Buxus sempervirens on rockslopes (Berberidion p.p.) (5110)Siliceous rocky slopes with chasmophytic vegetation (8220)Thermophilous Fraxinus angustifolia woods (91B0)Alluvial forests with Alnus glutinosa and Fraxinus excelsior (Alno-Padion,Alnion incanae, Salicion albae) (91E0*)Salix alba and Populus alba galleries (92A0)Quercus suber forests (9330)Quercus ilex and Quercus rotundifolia forests (9340)Endemic forests with Juniperus spp. (9560*)

Portuguese river valley system with Mediterranean type of climate and landscape mosaicdominated by open sclerophyllous woodland and scrub of Quercus suber, Quercus ilex and Juniperusoxycedrus, in mosaic with dry grasslands. In uplands and plateau areas oak forests, old chestnutgroves and dry scrub predominate in the landscape. Cliffs and steep slopes are common in themain valley as well as several types of riparian vegetation.Jackson, A.L.R., 2011. Renewable energy vs. biodiversity: policy conflicts and the future of natureconservation. Glob. Environ. Change 21, 1195–1208.

Peneda-Gerês PTZPE0002andPTCON0001

Temperate Atlantic wet heaths with Erica ciliaris and Erica tetralix (4020*)European dry heaths (4030)Arborescent matorral with Laurus nobilis (5230*)Oro-Iberian Festuca indigesta Grasslands (6160)Species-rich Nardus grasslands, on siliceous substrates in mountain areas(and submountain areas, in Continental Europe) (6230*)Lowland hay meadows (Alopecurus pratensis, Sanguisorba officinalis) (6510)Siliceous rocky slopes with chasmophytic vegetation (8220)Siliceous rock with pioneer vegetation of the Sedo-Scleranthion or of the Sedoalbi-Veronicion dillenii (8230)Alluvial forests with Alnus glutinosa and Fraxinus excelsior (Alno-Padion,Alnion incanae, Salicion albae) (91E0*)Galicio-Portuguese oak woods with Quercus robur and Quercus pyrenaica(9230)

Portuguese upland mosaic of mountain farmland and grazing lands, including native deciduousoak forests, hay meadows, acidophyllous grasslands and heathlands. Lowland valleys aredominated by pine plantations, oak woodlands and dry heath. Riparian forests and laurel thicketsare common along river valleys. Saxicolous habitats are common at high elevations and hostendemic plant species and vegetation types.Cerqueira, Y., Araújo, C., Vicente, J., Pereira, H.M., Honrado, J., 2010. Ecological and culturalconsequences of agricultural abandonment in the Peneda-Gerês National Park (Portugal). Chapter2 in: Evelpidou, N., Figueiredo, T., Mauro, F., Tecim, V., Vassilopoulos, A. (Eds.) “Natural Heritagefrom East to West”. 1st Edition, 2010, XIII, 363 p., Springer.

Cors Fochno UK0014791 Active raised bogs (7110)Degraded raised bog (7120)

One of the largest surviving lowland bogs in the UK with both an active raised bog and degradedraised bog within an estuarine setting.http://jncc.defra.gov.uk/ProtectedSites/SACselection/sac.asp?EUCode=UK0014791

Wekeromse Zand andEderheide & Ginkelseheide

NL9801023 Inland dunes with open Corynephorus and Agrostis grasslands (2330)Dry sand heaths with Calluna and Genista (2310), Eur. dry heaths (4030)Species-rich Nardus grasslands, on silicious substrates (6230)Atlantic acidophilous beech forests with Ilex (9120), Old acidophil. oakwoods with Q. robur on sandy plains (9190)

A part of Veluwe National Park, the largest end moraine in the Netherlands that was created duringpenultimate glacial period. The final landscape of alternating sand dune areas, heathlands and dryforests were created by a long history of intensive land use (starting from Celtic fields in Iron Age).Mücher, C.A., Kooistra, L., Vermeulen, M., Vanden Borre, J., Haest, B., Haveman, R., 2013.Quantifying structure and function of Natura 2000 heathland habitats using spectral mixtureanalysis and segmentation techniques on hyperspectral imagery. Ecol. Indic. 33, 71–81.

Murgia Alta IT9120007 Semi-natural dry grasslands and scrubland facies on calcareous substrates(Festuco-Brometalia) (important orchid sites) (6210) Pseudo-steppe withgrasses and annuals of the Thero-Brachypodietea (6220*)Eastern sub-Mediterranean dry grasslands (Scorzoneratalia villosae) (62A0)Calcareous rocky slopes with chasmophytic vegetation (8210)Eastern white oak woods (91AA)Quercus trojana woods (9250)

Italian semi-natural grasslands and other semi-natural vegetation (garrigues, scrubs, Quercuspubescens, Q. trojana and Q. coccifera coppices) and conifer plantations cast in an agriculturallandscape matrix dominated by cereal crops and, to a lesser extent, comprised of olive groves,almond orchards and vineyards.Mairota, P., Leronnii, V., Xi, W., Mladenoff, D.J., Nagendra, H., 2014. Using spatial simulations ofhabitat modification for adaptive management of protected areas: Mediterranean grasslandmodification by woody plant encroachment. Environ. Conserv. 41, 144–156.

H.

Nagendra

et al.

/ International

Journal of

Applied

Earth O

bservation and

Geoinform

ation 37

(2015) 124–132

127

Le Cesine IT9150032andIT9150014

Coastal lagoons (1150)Annual vegetation of drift lines (1210)Salicornia and other annuals colonizing mud and sand (1310)Mediterranean salt meadows (Juncetalia maritimi) (1410)Mediterranean and thermo-Atlantic halophilous scrubs (Sarcocorneteafruticosi) (1420)Embryonic shifting dunes (2110)Shifting dunes along the shoreline with Ammophila arenaria (‘white dunes’)(2120)Malcolmietalia dune grasslands (2230)Coastal dunes with Juniperus spp. (2250)Mediterranean temporary ponds (3170)Calcareous fens with Cladium mariscus and species of the Caricion davallianae(7210)

Italian complex of coastal lagoons with intricate coenological pattern. Fine grained mosaic ofhelophytic stands, stripes of halophilous and dry therophytic vegetation, water pools, marshes.Sand dunes banks host remnants of Juniperus macrocarpa subsp. macrocarpa communities. Inlandthis complex is surrounded by a forest landscape mosaic composed of Mediterranean maquis andgarrigues, Quercus ilex woods and Pinus halepensis plantations.Tomaselli, V., Sciandrello, S., Di Pietro, R., 2011. Plant communities structure and composition inthree coastal wetlands in southern Apulia (Italy). Biologia 66, 1027–1043.Tomaselli, V., Tenerelli, P., Sciandrello, S., 2012. Mapping and quantifying habitat fragmentation insmall coastal areas: a case study of three protected wetlands in Apulia (Italy). Environ. Monit.Assess. 184, 693–713.Tomaselli, V., Dimopoulos, P., Marangi, C., Kallimanis, A.S., Adamo, M., Tarantino, C., Panitsa, M.,Terzi, M., Veronico, G., Lovergine, F., Nagendra, H., Lucas, R.M., Mairota, P., Mücher, C.A., Blonda, P.,2013. Translating land cover/land use classifications to habitat taxonomies for landscapemonitoring: a Mediterranean assessment. Landsc. Ecol. 28, 905–930.

Kalamas River delta GR2120001 Coastal lagoons (1150)Salix alba and Populus alba galleries (92AO) andSouthern riparian galleries and thickets (Nerio-Tamaricetea and Securinegiontinctoriae) (92DO)Sarcopoterium spinosum phryganas (5420)Mediterranean salt meadows (Juncetalia maritimi) (1410)Mediterranean and thermo-Atlantic halophilous scrubs (Sarcocorneteafruticosi) (1420)

Landscape mosaic in Greece, comprised of riparian forest stands and a coastal lagoon as well assalt meadows, halophilous scrubs, reed and tamarisk thickets, estuaries, coastal lagoons, andcoastal dune systems.Goedecke, F., 2009. Diversität und Zustandsbewertung der Auwälder- und Gebüsche imFFH-Gebiet Kalamas, NW-Griechenland. In: Biologischen Fakultät – Albert-von-Haller Instituts fürPflanzenwissenschaften, Göttingen: Georg-August-Universität Göttingen, pp. 143.Tomaselli, V., Dimopoulos, P., Marangi, C., Kallimanis, A.S., Adamo, M., Tarantino, C., Panitsa, M,Terzi, M, Veronico, G, Lovergine, F, Nagendra, H, Lucas, RM, Mairota, P, Mücher, C.A., Blonda, P.,2013. Translating land cover/land use classifications to habitat taxonomies for landscapemonitoring: a Mediterranean assessment. Landsc. Ecol. 28, 905–930.

Kalodiki Fen GR2120002 Brackish marshes (Cladium mariscus) (7210)Sarcopoterium spinosum phryganas (5420)Hellenic forests with Quercus coccifera (934A)

Wetland in Greece consisting of two lakes embedded in a topogenous mire, and Kalodiki Fen, aninland, highly biodiverse freshwater wetland.Dimopoulos, P., Sykora, K., Gilissen, C., Wiecherink, D., 2005. Vegetation ecology of Kalodiki Fen(NW Greece). Biologia (Bratislava) 60, 69–82.Sarika-Hatzinikolaou, P., Koumpli-Sovantzi, M.L., Yannitsaros, A., 1996. The vascular flora of LakeKalodiki (Ipiros, NW Greece). Webbia 50, 223–236.Sarika-Hatzinikolaou, P., Koumpli-Sovantzi, M.L., Yannitsaros, A., 1996. The vascular flora of LakeKalodiki (Ipiros, NW Greece). Webbia 50, 223–236.Tomaselli, V., Dimopoulos, P., Marangi, C., Kallimanis, A.S., Adamo, M., Tarantino, C., Panitsa, M.,Terzi, M., Veronico, G., Lovergine, F., Nagendra, H., Lucas, R.M., Mairota, P., Mücher, C.A., Blonda, P.,2013. Translating land cover/land use classifications to habitat taxonomies for landscapemonitoring: a Mediterranean assessment. Landsc. Ecol. 28, 905–930.

Kalamas River Gorge GR2120004 Riparian forests with Platanus orientalis (92C0)Salix alba and Populus alba galleries (92A0)Sarcopoterium spinosum phryganas (5420)

Riparian forest landscape in Greece dominated by deciduous woodlands and riparian forest.Goedecke, F., 2009. Diversität und Zustandsbewertung der Auwälder- und Gebüsche imFFH-Gebiet Kalamas, NW-Griechenland. In: Biologischen Fakultät – Albert-von-Haller Instituts fürPflanzenwissenschaften. Göttingen: Georg-August-Universität Göttingen, pp. 143.Tomaselli, V., Dimopoulos, P., Marangi, C., Kallimanis, A.S., Adamo, M., Tarantino, C., Panitsa, M.,Terzi, M., Veronico, G., Lovergine, F., Nagendra, H., Lucas, R.M., Mairota, P., Mücher, C.A., Blonda, P.,2013. Translating land cover/land use classifications to habitat taxonomies for landscapemonitoring: a Mediterranean assessment. Landsc. Ecol. 28, 905–930.

Biligiri RangaswamyTemple TigerReserve†

N/A Evergreen forestsSemi-evergreen forestsShrub Savanna (grasslands)Riparian forestsShrub woodland to thickets with bambooTree Savanna

Highly biodiverse landscape located between the Eastern and Western Ghats region of India. Thearea surrounding the park is densely populated, with agriculture providing the main occupationand high dependence on natural vegetation for fuelwood and fodder.Ramesh, B.R., 1989. Flora of Biligiri Rangan Hills. Ph.D. thesis, Madras University.

(*) For European Union sites habitats types of Annex I of the European Habitats Directive (Council Directive 92/43/EEC) are indicated according to the European Commission, 2013. Interpretation Manual of EuropeanHabitats–EUR28. Published by the European Commission, DG Environment, Nature and Biodiversity.(†) Not a European Union site, thus Natura 2000 system and Annex I habitat type definitions do not apply.

1 arth O

fpbtpSoi

L

o2vkvpo(

wBto

L

ictbC2cetw(d(

brIrtumacst((

H

sht(ctmqt

28 H. Nagendra et al. / International Journal of Applied E

or classifications of habitat types and habitat conditions. Theserocesses, often fuzzy, scale dependent and site specific, need toe translated into explicit rules and algorithms that can be sequen-ially applied across locations, to convert EO data into analyticalroducts. The importance of expert interpretation is highlighted byhao and Wu (2008). The sections below provide a brief overviewf the approaches used by analysts for conversion of EO data intonformation on various categories of changes in state.

and cover/habitat conversionThe detection of land cover/habitat change frequently relies

n supervised processes of change detection (Tarantino et al.,007), requiring ground reference information, which can be pro-ided from field data, by photointerpretation or from prior expertnowledge of the site, for training the classifiers and subsequentlyalidating the results. Classification is typically conducted usinger-pixel classifiers, but object based change detection techniquesften provide better performance for very high spatial resolutionVHR) imagery (Chen et al., 2012).

Land cover/habitat conversions, once detected, can be relatedith information on changes in species diversity (Nagendra, 2001;uchanan et al., 2008). These can also be related to habitat fragmen-ation and changes in landscape connectivity, through the couplingf GIS and ecological modelling with EO.

and cover/habitat modificationIt is challenging to map land cover and habitat modification, for

nstance when these involve changes taking place below the topanopy in forest habitats (Joseph et al., 2011). Land cover mappinghrough VHR EO images and subsequent translations to habitats,ased on the Food and Agricultural Organization (FAO) Land Coverlassification System (LCCS) taxonomy (Di Gregorio and Jansen,005), have been demonstrated to be helpful for assessing landover and habitat modification (Lucas et al., this issue; Tomasellit al., 2013). As an example, habitat modification in native vegeta-ion can be detected by relating changes in the LCCS classifier cover,hich ranges from closed (up to 65%) to open (65–15%) and sparse

15–1%). VHR images are very useful to assess habitat modificationue to pressures such as pest infestation (Fuller, 2007) and grazingAllard, 2003) in diverse terrestrial habitats.

Direct associations have been established between spectralands or indices, and plant diversity (Joseph et al., 2011). Textu-al, spectral and structural features developed from EO (Koch andvits, 2004; Costanza et al., 2011) have also been related to speciesichness of plants, birds and insects. Changes in within-patch spa-ial patterns have provided indicators of change in habitat qualitysing indicators such as texture (St-Louis et al., 2006), and spectralixture analysis (Mücher et al., 2013). Active remote sensing such

s LiDAR, with its ability to penetrate through the top vegetationanopy, has been used to map changes in the three-dimensionaltructure of vegetation (Vierling et al., 2008), relating these changeso silvicultural practices (Pirotti, 2011), and to habitat modificationsGraf et al., 2009), sometimes in combination with VHR imageryAdamo et al., 2014).

abitat fragmentation and changes in landscape connectivityIndicators of habitat fragmentation and changes in landscape

patial pattern have been used to monitor changes in state due touman and natural pressure (Pôc as et al., 2011). Changes in spa-ial pattern can be related to habitat suitability for specific speciesTorres et al., 2010). EO derived land use/land cover (LULC) mapsan be very important to derive baseline characterization of spa-

ial pattern for monitoring (Mairota et al., 2013), and can assist in

odelling the ecological implications of fragmentation on habitatuality (Mairota et al., 2014a,b). Changes in spatial pattern observedhrough EO can also provide useful inputs to improve modelling of

bservation and Geoinformation 37 (2015) 124–132

species and community distribution (Padoa-Schioppa et al., 2012),of landscape functional connectivity (Mairota et al., 2011), and offuture vegetation dynamics (Mairota et al., 2014a,b).

Changes in plant community structureVHR imagery has been utilized for the delineation of top canopy

vegetation such as invasive plants (He et al., 2011), trees (Agarwalet al., 2013), reeds (Everitt et al., 2005), and grasses (Hall et al.,2011). Pressures that cause disruptions in plant community struc-ture are more difficult to assess using EO. The date of imagecollection is critical, and should ideally be timed to coincide withspecific phenological stages that make species of interest easy toidentify (Turner et al., 2003). For species mapping, there are impor-tant trade-offs between spatial and temporal resolutions of satelliteimagery, which are difficult to resolve (Nagendra and Rocchini,2008).

Results: applications of the approach

Utilizing the approach described above, and drawing on infor-mation from EO analysts familiar with the protected areas studiedin BIO SOS, a site and habitat specific categorization of changes instate and pressures is described (Supplement 1). Based on expertinformation from collaborations between ecologists and EO dataanalysts familiar with each site, we found that the majority of thechanges in state in the sites seem to be distinguishable throughEO. However certain types of changes in state, such as soil dis-turbance due to sod cutting in The Netherlands, or changes inhabitat quality due to point pollution sources in various sites, areunlikely to be detected using EO. Expert understanding and cross-talk between ecologists and EO analysts provides important inputsinto the process of target identification, EO data procurement, andidentification of the most appropriate methods of analysis. Acrosssites belonging to different biogeographic regions characterized bya range of habitat types and socio-ecological conditions, the spatialand temporal resolution of EO datasets should match the temporaland spatial scales of the pressures observed in each site (Supple-ment 1).

Two case studies reported as an extended demonstration of thisapproach illustrate how expert knowledge can be used to relatechanges in state with specific ongoing pressures in each protectedsite.

Case study – Le Cesine

Le Cesine (Natura 2000 SCI code IT9150032, SPA codeIT9150014) is a coastal area, mainly composed by a complex ofcoastal lagoons and characterized by a system of ponds, marshesand wet meadows. The site covers an area of about 21.5 km2

(2148 ha). The wetland is one of the most important sites in south-ern Italy in terms of biodiversity. Reed and sedge communitiesare prone to fires spreading from the adjacent agricultural areasand from arson. Outside the boundary of the protected area, agolf course was created in an area formerly covered by a mosaicof Mediterranean grassland (Annex I 6220* “Pseudo-steppe withgrasses and annuals of the Thero-Brachypodietea” of the 92/43 EECDirective–EUNIS E1.313 for the annual grasslands) and scrub habi-tats (EUNIS F6.2C). Part of the same area was subsequently modifiedto create artificial water bodies (tanks) to supply the golf green. EOdata was used to assess land cover conversion from grassland toartificial water bodies. Such an assessment requires the post clas-sification comparison of two land cover maps from different years

(Bruzzone and Bovolo, 2013). Multi-seasonal information (in thislocation, for spring and winter images) would have been usefulfor classification purposes given the low number of spectral bands(four) in QuickBird images. However, due to the limited archive

H. Nagendra et al. / International Journal of Applied Earth Observation and Geoinformation 37 (2015) 124–132 129

F ture aa

avas(fate

ig. 1. (a) Quickbird image at time T1; (b) Quickbird image at time T2; (c) NDVI feand T2; (f) thresholded NDVI image difference between time T1 and T2.

vailability for specific years of interest, a procedure of unsuper-ised direct image comparison (Bruzzone and Bovolo, 2013) wasdopted. The analysis utilized two VHR QuickBird images, at 2.4 mpatial resolution, acquired on July 2005 (Fig. 1a), and on June 2009Fig. 1b). Image-differencing was applied to the normalized dif-

erence vegetation index (NDVI) computed from the two images,fter relative radiometric calibration of each of the images, aimedo reduce false changes. A threshold was applied to the differ-nce image to detect areas of change. Visual interpretation of VHR

t time T1; (d) NDVI feature at time T2; (e) NDVI image difference between time T1

images by expert ecologists working in combination with EO ana-lysts was essential to relate changes observed in the images withland cover class conversion. Thus, the pressure active in this regionwas related to the presence of a golf course. Besides the direct lossof important sections of the grassland habitat at the margins of

the protected area, such pressures can ultimately result in indi-rect impacts on aquatic habitat types located in the protected area,through changes in water-table salinity induced by surface waterextractions from the tanks.

130 H. Nagendra et al. / International Journal of Applied Earth Observation and Geoinformation 37 (2015) 124–132

F ue ratt

C

PSNhio

ig. 2. (a) Worldview image at time T1; (b) worldview image at time T2; (c) red/blime T1 and T2; (f) thresholded red/blue image difference between time T1 and T2.

ase study – Ederheide and Ginkelse heide

Ederheide and Ginkelse heide are located within the Nationalark “Veluwe” in the Province of Gelderland (SCI code NL9801023,PA code NL3009017). The Veluwe Natura 2000 site in The

etherlands (Fig. 2) consists of sand dune areas alternating witheathlands and dry forests, and is shaped by a long history of

ntensive land use. Ederheide and Ginkelse heide cover an areaf approximately 1000 ha, where heathland is the dominant land

io at time T1; (d) red/blue ratio at time T2; (e) red/blue image difference between

cover. Urbanization, tourism and agriculture are the main pressuresaffecting the site.

Inland sand dunes in this region have been impacted from nitro-gen deposition caused by intensive agriculture outside the park,leading to encroachment of moss, grass, shrubs and trees. The goal

of the analysis was to detect land cover changes associated with cut-ting of forest plantation and sod, implemented by park managers,to restore sand dune dynamics that were impacted due to excessivenitrogen deposition. The analysis utilized two Worldview II images,

arth O

a2tmwEew

D

as2irnata

acbtsc(altbtcwcwatwascn

bopcaaSti

caplWrp(e

H. Nagendra et al. / International Journal of Applied E

t 2 m spatial resolution, acquired on June 2011 (Fig. 2a) and March013 (Fig. 2b) respectively (Mücher et al., this issue). The approachaken was similar to the previous study. Following relative radio-

etric calibration, a red/blue band image-differencing algorithmas applied to the two images to identify locations of change.

xpert knowledge was used to relate these changes to cutting of for-st plantation and sod covered areas, and to associate these changesith direct interventions by protected area managers.

iscussion

The usefulness of EO data to inform biodiversity conservationnd environmental management has been highlighted by remoteensing specialists and ecologists in recent years (Nagendra et al.,013; Pettorelli et al., 2014). The progress of such integration has,

t seems, been slower than anticipated (Pettorelli et al., 2014), andequires greater coordination between these two research commu-ities to develop a common shared understanding, terminology,nd problem definition. Such synchronization is key to improvinghe potential for EO data to effectively support environmental man-gement decisions by providing adequate value added products.

The paper builds on collaborations between ecologists and EOnalysts familiar with site-specific habitat and land cover relatedhanges in state in specific protected areas, to elicit the linkageetween the ecological concept of pressures and the EO detec-ion of change in state. While high resolution (HR) Landsat dataeries are available for free and can be used to extract specificlass trends through post-classification comparison of LULC mapsTarantino et al., 2007), change detection is frequently requestedt fine resolution for use in local/regional decision making. Theack of multi-seasonal very high resolution (VHR) archive data onhe same area can limit the production of accurate LULC maps toe compared for the detection of from-to class transitions, withhese closely related to pressures on the environment. In suchases, automatable, unsupervised techniques can be combinedith expert knowledge and visual interpretation of images to infer

lass transitions, as demonstrated in the two case studies presentedithin this paper. Some of the applied EO data analysis routines

re well established. The importance lies in the close collabora-ion between ecologists and EO analysis, both of whom are familiarith the site specific context of pressures and changes in state in the

reas of study. This enables the selection of appropriate site-specificpatial and temporal scales of EO data for identifying pressures andhanges in state, and of identification of best EO analytical tech-iques.

VHR data are only tasked on specific areas and requirementsut not collected globally, as HR Landsat data is for example. Usef different VHR sensors for comparison of land cover change isroblematic due to leads to variation in critical factors such aslassification accuracy, spatial resolution, data processing, meanccuracy, and seasonality (Shao and Wu, 2008). However, with thedvent of new imagery such as that provided by the launch of ESA’sENTINEL, open VHR data will become available with improvedemporal resolution, providing new opportunities for ecologicalnsights to be incorporated into EO analysis and monitoring.

In order to address the widespread demand of detecting VHRhanges in biodiversity state within and outside protected areas,greements need to be established between space agencies andrivate companies (such as DigitalGlobeTM, which provides the

argest sub-metre constellation of satellites, including QuickBird,orldView-1, WorldView-2, IKONOS and GeoEye-1) to enable the

egular acquisition of multi-seasonal images. Such agreements, asroduced in the past within the European Framework Programmee.g. DataWarehouse) etc., can greatly advance the provision ofcologically annotated EO monitoring data on changes in state to

bservation and Geoinformation 37 (2015) 124–132 131

provide support to national or regional conservation authoritiesand research bodies.

Selection of site-specific appropriate EO images and processingtechniques by EO analysts should be based on prior, expertsite-specific knowledge from ecologists. This should include infor-mation on plant physiology and phenology, vegetation structure,spatial relationships and temporal dynamics, and habitat andlandscape level disturbance regimes that are of use for discrim-inating land cover/habitat classes with similar signatures. Thisis the main recommendation of a joint white paper by FP-7SPACE projects BIO SOS (http://www.biosos.eu/) and MS.MONINA(http://www.ms-monina.eu/).

Based on expert judgement, the approach suggested and docu-mented in this paper (elaborated in the case studies) is applicableto selected protected areas in diverse biogeographic regions andwith a diversity of land cover/habitat types. This approach con-nects to the growing data sets from remote sources and patternrecognition techniques for their automatic interpretation. We urgeit to be tested and applied over a wide range of pressures andsocial–ecological settings.

Acknowledgments

This research draws on the European Union’s Seventh Frame-work Programme (EU-FP7) project titled Biodiversity Multi-SOurceMonitoring System: From Space To Species (BIO SOS, GA 263435)coordinated by CNR ISSIA, Bari, Italy (http://www.biosos.eu).We thank the anonymous reviewers of earlier versions of thismanuscript for their very useful feedback.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jag.2014.10.010.

References

Adamo, M., Tarantino, C., Tomaselli, V., Kosmidou, V., Petrou, Z., Manakos, I., Lucas,R.M., Mücher, C.A., Veronico, G., Marangi, C., De Pasquale, V., Blonda, P., 2014.Expert knowledge for translating land cover/use maps to general habitat cate-gories (GHC). Landsc. Ecol., http://dx.doi.org/10.1007/s10980-014-0028-9.

Agarwal, S., Vailshery, L.S., Jaganmohan, M., Nagendra, H., 2013. Mapping urban treespecies using very high resolution satellite imagery: comparing pixel-based andobject-based approaches. ISPRS Int. J. Geo-Inf. 2, 220–236.

Allard, A., 2003. Detection of vegetation degradation on Swedish mountainousheaths at an early stage by image interpretation. Ambio 32, 510–519.

Buchanan, G.M., Butchart, S.H.M., Dutson, G., Pilgrim, J.D., Steininger, M.K., Bishop,K.D., Mayaux, P., 2008. Using remote sensing to inform conservation statusassessment: estimates of recent deforestation rates on New Britain and theimpacts upon endemic birds. Biol. Conserv. 141, 56–66.

Bruzzone, L., Bovolo, F., 2013. A novel framework for the design of change-detectionsystems for very-high-resolution remote sensing images. Proc. IEEE 101 (3),609–630.

Bunce, R.G.H., Bogers, M.M.B., Evans, D., Halada, L., Jongman, R.H.G., Mücher, C.A.,Bauch, B., de Blust, G., Parr, T.W., Olsvig-Whittaker, L., 2013. The significance ofhabitats as indicators of biodiversity and their links to species. Ecol. Indic. 33,19–25.

Chen, G., Hay, G.J., Carvalho, L.M.T., Wulder, M.A., 2012. Object-based change detec-tion. Int. J. Remote Sens. 33, 4434–4457.

Costanza, J.K., Moody, A., Peet, R.K., 2011. Multi-scale environmental heterogeneityas a predictor of plant species richness. Landsc. Ecol. 26, 851–864.

Di Gregorio, A., Jansen, L.J.M., 2005. Land Cover Classification System (LCCS): Clas-sification Concepts and User Manual. Food and Agriculture Organization of theUnited Nations, Rome.

EEA, 1995. Europe’s Environment: The Dobris Assessment. European EnvironmentAgency, Copenhagen.

Everitt, J.H., Yang, C., Deloach Jr., C.J., 2005. Remote sensing of giant reed with Quick-Bird satellite imagery. J. Aquat. Plant Manage. 43, 81–85.

Fuller, D.O., 2007. Tropical forest monitoring and remote sensing: a new era oftransparency in forest governance? Singap. J. Trop. Geogr. 27, 15–29.

Graf, R.F., Mathys, L., Bollmann, K., 2009. Habitat assessment for forest dwellingspecies using LiDAR remote sensing: capercaillie in the Alps. For. Ecol. Manage.257, 160–167.

1 arth O

H

H

H

H

H

J

J

K

M

M

M

M

M

M

N

N

N

32 H. Nagendra et al. / International Journal of Applied E

all, K., Reitalu, T., Sykes, M.T., Prentice, H.C., 2011. Spatial heterogeneity of Quick-Bird satellite data is related to fine-scale plant species spatial turnover insemi-natural grasslands. Appl. Veg. Sci. 15, 145–157.

e, K.S., Rocchini, D., Neteler, M., Nagendra, H., 2011. Benefits of hyper-spectral remote sensing for tracking plant invasions. Divers. Distrib. 17,381–392.

ockings, M., Stolton, S., Leverington, F., Dudley, N., Courrau, J., 2006. EvaluatingEffectiveness. A Framework for Assessing Management Effectiveness of Pro-tected Areas, second ed. IUCN, Gland, Switzerland/Cambridge, UK.

onrado, J.P., Viera, C., Soares, C., Monteiro, M.B., Marcos, B., Pereira, H.M., Partidário,M.R., 2013. Can we infer about ecosystem services from EIA and SEA practice?A framework for analysis and examples from Portugal. Environ. Impact Assess.40, 14–24.

ooper, D.U., Adair, E.C., Cardinale, B.J., Byrnes, J.E.K., Hungate, B.A., Matulich, K.L.,Gonzalez, A., Duffy, J.E., Gamfeldt, L., O’Connor, M.I., 2012. A global synthesisreveals biodiversity loss as a major driver of ecosystem change. Nature 486,105–108.

enkins, C.N., Joppa, L., 2009. Expansion of the global terrestrial protected area sys-tem. Biol. Conserv. 142, 2166–2174.

oseph, S., Murthy, M.S.R., Thomas, A.P., 2011. The progress on remote sensing tech-nology in identifying tropical forest degradation: a synthesis of the presentknowledge and future perspectives. Environ. Earth Sci. 64, 731–741.

och, B., Ivits, E., 2004. Results from the Project BIOASSESS – relation betweenremote sensing and terrestrial derived biodiversity indicators. In: Monitoringand Indicators of Forest Biodiversity in Europe–From Ideas to Operationality EFIProceedings No. 51, pp. 315–332.

ace, G., Norris, K., Fitter, A.H., 2012. Biodiversity and ecosystem services: a multi-layered relationship. Trends Ecol. Evol. 27, 19–26.

airota, P., Leronni, V., Cafarelli, B., Baveco, J.M., 2011. Connectivity in a real frag-mented landscape: distance vs. movement model based approaches. In: SpatialData Methods for Environmental and Ecological Processes, second ed. CDPService edizioni, Foggia, pp. 145–148, ISBN: 9788896025123.

airota, P., Cafarelli, B., Boccaccio, L., Leronni, V., Labadessa, R., Kosmidou, V., Nagen-dra, H., 2013. Using landscape structure to develop quantitative baselines forprotected area monitoring. Ecol. Indic. 33, 82–95.

airota, P., Cafarelli, B., Labadessa, R., Lovergine, F.P., Tarantino, C., Nagendra, H.,Didham, R.K., 2014a. Very high resolution Earth observation features for testingthe direct and indirect effects of landscape structure on local habitat quality. Int.J. Appl. Earth Observ. Geoinf., http://dx.doi.org/10.1016/j.jag.2014.07.003

airota, P., Leronni, V., Xi, W., Mladenoff, D., Nagendra, H., 2014b. Using spatial sim-ulations of habitat modification for adaptive management of protected areas:Mediterranean grassland modification by woody plant encroachment. Environ.Conserv. 41, 144–156.

ücher, C.A., Kooistra, L., Vermeulen, M., Vanden Borre, J., Haest, B., Haveman, R.,2013. Quantifying structure and function of Natura 2000 heathland habitatsusing spectral mixture analysis and segmentation techniques on hyperspectralimagery. Ecol. Indic. 33, 71–81.

agendra, H., 2001. Using remote sensing to assess biodiversity. Int. J. Remote Sens.22, 2377–2400.

agendra, H., Rocchini, D., 2008. High resolution satellite imagery for tropi-cal biodiversity studies: the devil is in the detail. Biodivers. Conserv. 17,3431–3442.

agendra, H., 2008. Do parks work? Impact of protected areas on land cover clearing.Ambio 37, 330–337.

bservation and Geoinformation 37 (2015) 124–132

Nagendra, H., Rocchini, D., Ghate, R., 2010. Beyond parks as monoliths: spatiallydifferentiating park-people relationships in the Tadoba Andhari Tiger Reservein India. Biol. Conserv. 143, 2900–2908.

Nagendra, H., Lucas, R., Honrado, J.P., Jongman, R.H.G., Tarantino, C., Adamo, M.,Mairota, P., 2013. Remote sensing for conservation monitoring: assessing pro-tected areas, habitat extent, habitat condition, species diversity and threats. Ecol.Indic. 33, 45–59.

Nagendra, H., Mairota, P., Blonda, P., Marangi, C., Torri, D., Lucas, R., Dimopolous,P., Honrado, J.P., Niphadkar, M., Mücher, S., Tomaselli, V., 2012. Methodologyto Identify and Quantify Local Pressures. Biodiversity Multisource MonitoringSystem: From Space TO Species (BIO SOS) Deliverable D6.8.

Padoa-Schioppa, E., Ficetola, F.G., Bonardi, A., Mücher, C.A., Nipadkar, M., Nagendra,H., Leronni, V., Mairota, P., Tarantino, C., Adamo, M., 2012. Report on EcologicalNiche Modeling Biodiversity Multisource Monitoring System: From Space TOSpecies (BIO SOS) Deliverable D6.7.

Pettorelli, N., Laurance, W.F., O’Brien, T., Wegmann, M., Nagendra, H., Turner, W.,2014. Satellite remote sensing for applied ecologists: opportunities and chal-lenges. J. Appl. Ecol. 51, 839–848.

Pirotti, F., 2011. Analysis of full-waveform LiDAR data for forestry applications: areview of investigations and methods. iForest Biogeosci. For. 4, 100–106.

Pôc as, I., Cunha, M., Pereira, L.S., 2011. Remote sensing based indicators of changesin a mountain rural landscape of Northeast Portugal. Appl. Geogr. 31, 871–880.

Salafsky, N., Salzer, D., Stattersfield, A.J., Hilton-Taylor, C., Neugarten, R., Butchart,S.H.M., Collen, B., Cox, N., Master, L.L., O’Connor, S., Wilkie, D., 2008. A standardlexicon for biodiversity: unified classifications of threats and actions. Conserv.Biol. 22, 897–911.

Shao, G., Wu, J., 2008. On the accuracy of landscape pattern analysis using remotesensing data. Landsc. Ecol. 23, 505–511.

St-Louis, V., Pidgeon, A.M., Radeloff, V.C., Hawbaker, T.J., Clayton, M.K., 2006. High-resolution image texture as a predictor of bird species richness. Remote Sens.Environ. 105, 299–312.

Tarantino, C., Blonda, P., Pasquariello, G., 2007. Remote sensed data for automaticdetection of land-use changes due to human activity in support to landslidestudies. Nat. Hazards 41, 245–267.

Tomaselli, V., Dimopoulos, P., Marangi, C., Kallimanis, A.S., Adamo, M., Tarantino, C.,Panitsa, M., Terzi, M., Veronico, G., Lovergine, F., Nagendra, H., Lucas, R., Mairota,P., Mücher, S., Blonda, P., 2013. Translating land cover/land use classificationsto habitat taxonomies for landscape monitoring: a Mediterranean assessment.Landsc. Ecol. 28, 905–930.

Torres, J., Brito, J.C., Vasconcelos, M.J., Catarino, L., Gonc alves, J., Honrado, J., 2010.Ensemble models of habitat suitability relate chimpanzee (Pan troglodytes) con-servation to forest and landscape dynamics in Western Africa. Biol. Conserv. 143,416–425.

Turner, W., Spector, S., Gardiner, N., Fladeland, M., Sterling, E., Steininger, M., 2003.Remote sensing for biodiversity science and conservation. Trends Ecol. Evol. 18,306–314.

Vicente, J.R., Fernandes, R.F., Randin, C.F., Broennimann, O., Gonc alves, J., Marcos,B., Pôc as, I., Alves, P., Guisan, A., Honrado, J.P., 2013. Will climate change drivealien invasive plants into areas of high conservation value? An improved model-

based regional assessment to prioritize the management of invasions. J. Environ.Manage. 131, 185–195.

Vierling, K.T., Vierling, L.A., Gould, W.A., Martinuzzi, S., Clawges, R.M., 2008. Lidar:shedding new light on habitat characterization and modeling. Front. Ecol. Envi-ron. 6, 90–98.