Ecological Indicators 37 (2014) 163– 174
Sa
DI
a
KRPSP
1
toshSoal(adp
1h
Contents lists available at ScienceDirect
Ecological Indicators
jou rn al hom epage: www.elsev ier .com/ locate /eco l ind
patial quantification and valuation of cultural ecosystem services in angricultural landscape
erek B. van Berkel ∗, Peter H. Verburgnstitute for Environmental Studies (IVM) and Amsterdam Global Change Institute, VU University, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands
r t i c l e i n f o
eywords:ural developmentarticipatory Mappingtated and Revealed monetary valuationhoto-realistic montages
a b s t r a c t
While the spatial and economic quantification and valuation of ecosystem services is becoming increas-ingly recognised as a way to communicate the importance of ecosystem conservation, little attentionhas been given to cultural services of the landscape. Cultural services form an important part oftourism amenities in agricultural landscapes. In this study we present a methodology for quantify-ing cultural services. To gain understanding of the services valued by cultural service users, a surveywas conducted with tourists in the municipality of Winterswijk. The survey collected data on land-scape preferences for individual landscape features, and the structure and composition of the landscapeas a whole. This was linked to respondent appreciation of the landscape functions of recreation,aesthetic beauty, cultural heritage, spirituality and inspiration. To give a monetary estimate of thevalue of these services a willingness to pay (WTP) exercise was conducted using photo manipula-tions depicting likely landscape changes. Increased residential infill, the removal of landscape elementsfor improved agricultural production and rewilding due to agricultural abandonment were simulated.Complementary to this estimate, a travel cost estimate of the value of landscape service was done
based on respondents’ travel time to reach the region. The monetary value of the cultural servicesis placed between D 86 (WTP) and D 23 (travel cost) per tourist/year. The achieved understanding ofthe spatial heterogeneity of service provision in the region, as well as, the monetary valuation of theassets delivered by the landscape can help in prioritizing areas, and landscape features and struc-ture for maintenance/restoration, while demonstrating the importance of conserving cultural servicedelivery.© 2012 Elsevier Ltd. All rights reserved.
. Introduction
Humans benefit from the numerous services that rural ecosys-ems deliver whether that is the provision of food, the regulationf clean water or the inspiration invoked by a beautiful land-cape (MA, 2003). In Europe, many agricultural landscapes areot spots of ecosystem service delivery (Pinto-Correia et al., 2006;olymosi, 2011; Stenseke, 2009). Such agricultural landscapes areften denoted as cultural landscapes, which are typically defineds landscapes managed by traditional agricultural techniques,ocally adapted and historic, by family and/or subsistence methodsIEEP, 2007). Often they contribute to a unique aesthetic char-cter and support a co-produced human–ecological system. Yet,
ue to processes of agricultural intensification, occurring in manyarts of Europe, cultural landscape are being transformed in ways∗ Corresponding author. Tel.: +31 205989556.E-mail address: [email protected] (D.B. van Berkel).
470-160X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.ecolind.2012.06.025
that negatively affect the delivery of cultural ecosystem services(Zimmermann, 2006).
Over the last decades there has been much attention givento maintaining spatial and economic synergies between ecosys-tem functions in rural areas as part of development planning. Thisis generally thought to allow local communities to better copewith the various endogenous and exogenous pressures that canthreaten livelihoods in these landscapes (Marsden and Sonnino,2008; Knickel et al., 2004; O’Farrell and Anderson, 2010; Rentinget al., 2009; Wilson, 2010). Promotion of tourism and recreation,based on the existing features and traditions, is a preferred ruraldevelopment option (Van Berkel and Verburg, 2011). It enablesincome generation outside of agricultural production intensifica-tion and promotes the preservation of existing assets (Buijs et al.,2006; Marsden, 1999). Tourism attractions are related to people’sawareness and perceived importance of aesthetic beauty, cultural
heritage, spirituality and inspiration (Brown, 2006). Such charac-teristics are non-material benefits related to land management andtherefore non-exclusive. Failure to provide enough incentives forthe maintenance of cultural landscapes may result in their loss
1 ologic
atsd
bSi2eirthsttSttRehdv
diedHSawthlvt2iei
sWsaasvo
Noaug1ispmtre
(forest, tree lines, recreation facilities, cultural buildings, etc.)
64 D.B. van Berkel, P.H. Verburg / Ec
nd/or degradation (Swinton et al., 2007). The quantification ofhe cultural services provided by landscapes both in monetary andpatial terms can contribute to understanding options for futureevelopment that retain tourism assets.
Major contributions have been made to the understanding ofoth the monetary costs and benefits of ecosystem service delivery.tudies mapping ecosystem services have offered policymakersnsight into priority locations for service delivery (Egoh et al.,008; Lautenbach et al., 2011; Nedkov and Burkhard, 2011; Nelsont al., 2009; Willemen et al., 2008). These studies are often lim-ted to examining provisioning and regulatory services based oneadily available biophysical data. The normative nature of cul-ural services and the heterogeneity in valuation of societal actorsas made their quantification more difficult (Ryan, 2011). Mosttudies evaluating ecosystem services have been limited to quan-ifying recreation and tourism, leaving out the intrinsic qualitieshat are interrelated with tourism in the cultural service category.till, a number of techniques have been developed for the localisa-ion of services valued by stakeholders, including cultural services,hrough participatory mapping (Alessa et al., 2008; Brown andaymond, 2007; Bryan et al., 2010; Dramstad et al., 2006; Raymondt al., 2009; Sherrouse et al., 2011). The identification of locations ofigh service delivery has been helpful for understanding the spatialeterminants of fortuitous ecosystem delivery, and its associatedalue to society.
One particular challenge for participatory mapping has beenescribing the monetary value of the identified services, which
s the focus of economic valuation of ecosystems. Revealed pref-rence techniques have been useful in estimating the actual andirect uses cost incurred by service users (Geoghegan et al., 1997;ein, 2011; Ma and Swinton, 2011; Martín-López et al., 2009;antana-Jimènez et al., 2011). While based on a number of broadssumptions, such techniques avoid respondent bias for instanceith warm glow responses (Hanley et al., 2001). Stated preference
echniques, including contingent valuation and discrete choice,ave been more widely used for valuations of non-use services
ike biodiversity (Birol et al., 2008). Such studies reveal the societalalues placed upon intrinsic characteristics while perhaps overes-imating the actual costs that individuals would pay (Hanley et al.,001). While debates abound regarding the accuracy and reliabil-
ty of derived prices, results have had major policy impact wherecosystem goods and service are now being considered seriouslyn ecosystem management (Kinzig et al., 2011).
This study adds to this body of literature by integrating both apatial quantification and economic valuation of cultural services.
e consider both individual landscape features and landscapetructure. This is then related to tourist experience and appreci-tion of recreation, aesthetic beauty, cultural heritage, spiritualitynd inspiration in the landscape. By characterizing preferences oftakeholders, a spatial localisation and analysis of landscape ser-ices is made. In addition, monetary valuation gives an indicationf how important these services are for the regional economy itself.
The research is conducted in the Achterhoek region of theetherlands, which has a well developed tourism industry basedn the cultural landscape and nature attractions. The easternreas have retained much of their preindustrial character due tonique historical circumstances that prevented farmers from reor-anising small parcels into large agricultural plots (Wildenbeest,989). The landscape is presently characterized by a network of
nterlinking tree lines and hedgerows called the coulissen land-cape. Tree shadows created by tree lines reduce agriculturalroduction and are a hindrance for modern farming equip-ent. This in conjunction with an aging farmer population and
he price production squeeze has resulted in some landownersemoving landscape elements for agricultural production scalenlargement.
al Indicators 37 (2014) 163– 174
2. Methodology
2.1. Method overview
The main aim of the study is to locate and quantify the cul-tural services provided by the landscape and provide a monetaryvaluation of these services. A differentiation of the contributionof individual elements of the landscape and the landscape com-position and structure to the provision of these services is made.Empirical data was collected in the eastern most municipalities ofthe Achterhoek (Fig. 1) by way of a questionnaire survey in the sum-mer of 2011. Statistical analysis was employed to identify groupsof respondents with similar appreciation of landscape functioningand to ascertain their preference for landscape features, structureand evolution. Preferences were then translated into maps showinghot spots of cultural service provision. Respondents’ willingness topay (WTP) for landscape maintenance is provided to give an esti-mate of the potential value of landscape services in the region,under conditions of ongoing change. A travel time/cost estimateis made of the revealed value of these landscape services to com-pliment the WTP estimate.
2.2. Survey
The questionnaire was administered in the municipality ofWinterswijk by an experienced survey team. Respondents wereinterviewed in person at campsites, agri-campsites, recreationareas (lakes, nature areas and popular tourist locations) in both theDutch and German language. This allowed for targeting the major-ity of tourists in different locations that contribute to the touristfunction of the region. The face to face survey method increasedresponse rates as compared with mail-in surveys which are diffi-cult to administer with tourists that do not reside in the region. Intotal 115 respondents took part in the survey. The average age ofthe sample was 53 with many visitors nearing retirement age orretired (50% older than 55). The average net income per respon-dent’s household was near the Dutch national average of 2315Dper month. The mean educational attainment was preparatory andsecondary vocational education (MBO, HBO). The sample group wascomprised of both ‘recreants’ and ‘tourists’. Recreants are defined asthose respondents living within a half an hour of their leisure activ-ity (n = 17) and tourists are all those living further away (n = 98). Theaverage travel time to reach Winterswijk was 1 h 23 min, which isapproximately the time needed to reach the destination from thecentral part of the Netherlands. The total sample size is compara-ble to other ecosystem service mapping studies (Bryan et al., 2010;Dramstad et al., 2006) while being smaller than national preferencesurveys employing mail-in questionnaires (Brouwer and Slangen,1998; Soliva et al., 2010).
2.3. Survey method
The questionnaire consisted of three parts: (1) personal datawas collected for analysis of the sample group and application ofthe travel time/cost method; (2) respondents’ appreciation for dif-ferent landscape features, structure and landscape changes weretaken; and (3) a monetary valuation of the current landscapewas estimated by asking respondents their WTP for landscapepreservation considering likely landscape changes. Preferenceswere obtained through respondents’ evaluations of photos andphoto manipulations (Figs. 2 and 3). Photos of individual land-scape elements representing different local landscape features
and aerial photos of landscape structure and composition wereused (representing different amounts and configuration of agricul-tural, forest and hedgerows/tree lines). A number of studies have
D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174 165
Fig. 1. Map of the study area.
Fig. 2. Respondents’ assessment of important landscape features for the Achterhoek region.
166 D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174
tructu
sc(2l(PcTtsitde(ts
2
ifaialib2omlo
Fig. 3. Respondents preference for landscape s
uccessfully employed photos comparing landscape changes forollecting empirical data about landscape aesthetic preferenceHowley et al., 2012; Mari Sundli, 2009; Ode et al., 2009; Soliva et al.,010) as well as, more abstract notions like preferences for different
andscape and land use developments using photo manipulationsSoliva et al., 2008; Tress and Tress, 2003; Van Berkel et al., 2011).hoto-realism can produce accurate responses due to image pre-ision and believability (Dockerty et al., 2005; Lovett et al., 2010).his realism was thought to trigger emotional reactions regardinghe intrinsic qualities that respondents associate with the land-cape. The photos were collected from Google EarthTM and throughnfield documentation allowing us to represent a diversity of loca-ions in the study area. To ensure that visual aesthetic preferenceid not play a role in respondent evaluation of the images, differ-nt photos of the same feature or landscape change were shownFigs. 2 and 4). Respondents were asked to choose from the tophree preferred and important landscape features and best land-cape structure depicted in the photos.
.4. Mapping cultural services
Two maps were created for mapping cultural service provision-ng. One was based on the preference for individual landscapeeatures, the other based on preferences of landscape structurend composition. Preferences for the landscape were translatednto maps by allocating the number of respondents that chose
particular feature or structure depicted in the photos to mapayers representing them. Participatory mapping has become anncreasingly popular way to identify locations valued by society foretter informing planners and policymakers (Brown and Raymond,007; Dramstad et al., 2006). Such techniques have made use
f maps where respondents can place and draw directly on theap non-monetary values for indicating important service deliveryocations. Aggregating individual assessments gives an indicationf the societal importance for location specific services rather than
re and composition in the Achterhoek region.
an accurate monetary valuation (Alessa et al., 2008; Brown andRaymond, 2007; Sherrouse et al., 2011). Unlike these techniqueswe translate respondents’ evaluation of photos to mapped layers.Empirical data collection using maps was considered unsuitablegiven that tourists may or may not be familiar with the geogra-phy of the region. Different map layers representing the featuresand structure depicted in the photos were collected from variousprovincial databases (Provincie Gelderland, 2010).
For the feature preference map, respondents’ preferences wereallocated to the locations indicated when the landscape elementsshown in the photo occur in the map (Fig. 2). Preferences for cul-tural buildings, recreation areas, landscape elements (25 m bufferaround tree lines and hedgerows), forests, marshes and other landcover types were allocated separately and an aggregate sum cal-culated to determine hotspots of total preference. For the feature‘animals’ spatial localisation was more difficult as data was limitedto the habitat range of iconic animal species (Water fowl, meadowbutterflies, bats and ring snakes). To approximate these preferencesa viewshed calculation, from biking and walking paths to the loca-tions of these habitats, indicating where there was a possibilityto view wildlife, was made. Sight lines (180◦) were calculated forthe horizon to the surrounding countryside with barriers like treelines and forest determining the view extent. The observer heightwas assumed to be 1.5 m given an average of bikers’ and walkers’view. The viewable area was categorised as high, low and mediumaccording to the number of observer points where the habitat rangecould be seen.
For the landscape structure map, the aerial photos used in thequestionnaire were analysed to determine the proportion of agri-culture, tree lines and forest depicted. This was done in Photoshopby classifying and calculating the amount of these structures. For
instance, the ‘forest photo’ was composed of 79% forest, 21% agri-cultural land and 0% of tree lines. To translate the proportion ofthese elements depicted in the photos to mapped layers, a neigh-bourhood calculation was done using separate maps of tree lines,
D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174 167
F f (i) rea
fwAltTtaWwl
birurdnawlantii
ig. 4. Case study photos and photo manipulations depicting landscape processes ogricultural production and (iii) rewildng of extensive pasture lands.
orests and agricultural land cover. The neighbourhood dimensionsere the same as the extent of the photos to ensure comparability.
fuzzy membership calculation was then conducted on resultingayers to represent how close each layer fit to the proportions ofhese landscape structures to that which was depicted in photos.he separate layers were combined to create maps representinghe 6 different photos that respondents could choose from. Finally,
weighted overlay was made to combine these different maps.eights were determined based on the amount of respondentsho preferred the individual photos illustrating areas of preferred
andscape structure and composition (Fig. 3).To assess the role of the individual cultural services provided
y the landscape respondents were asked to rate the landscapen terms of recreation, aesthetic beauty, cultural heritage, inspi-ation and spirituality on a five point Likert scale (with 1 beingnimportant and 5 really important). It was hypothesised thatespondents with different landscape appreciation would preferifferent structures and features. To test this, a principle compo-ent analysis (PCA) was done to identify groups of similar landscapeppreciation. A PCA was chosen as it was expected that respondentsould value multiple cultural services provided by the agricultural
andscapes. Examination of individual responses (eigen values)llowed for determining this overlapping appreciation, which was
ot possible with other clustering techniques. These groups werehen analysed to ascertain unique preference using an ANOVAndependent t-test. This is a common approach for determin-ng socio-economic and political factors contributing to landscapesidential infill in the landscape, (ii) removal of landscape element due to intensive
preferences (Philip, 1984; Soliva et al., 2010; Van den Berg andKoole, 2006). For instance, preference has been linked to gender andage categories (Soini et al., 2012) urban–rural differences (Howleyet al., 2012) and cultural background (Soliva et al., 2010). However,there are no studies known to the authors assessing intrinsic appre-ciation as a determinant for landscape preferences. Maps of culturalservice appreciation were developed by applying the preferences ofthese subgroups to the different layers representing different land-scape features and structures in a similar way as for the lumpedsurvey results.
2.5. Landscape monetary valuation
2.5.1. Monetary valuation techniquesA number of methods are available for estimating the mone-
tary value of environmental and cultural services including bothstated (willingness to pay: WTP) and revealed preference tech-niques. Techniques for estimating WTP include discrete choiceand contingent valuation (Swinton et al., 2007). In discrete choiceexperiments respondents are asked to compare different optionsof services delivery given the costs of policy intervention. Servicesare described and often visualised on cards where respondentscan choose service delivery according to the various prices indi-
cated (Campbell, 2007). Contingent valuation uses scenarios thatdescribe a threatened service provision requiring policy interven-tions to ensure continued delivery (Brouwer and Slangen, 1998;Colombo et al., 2006; Ready et al., 1997). Respondents are asked
1 ologic
wvciornintp2sdict(
sbttpttt
2
pwstitticitadttt
gcdlttepsodtatt
2
a
68 D.B. van Berkel, P.H. Verburg / Ec
hat amount of money they would be willing to pay for inter-entions to maintain or enhance service delivery. While discretehoice experiments appear to more accurately determine WTPn comparison to stated valuation, discrete choice exercises areften cognitively strenuous in terms of choice options and timeequirements (Hanley et al., 2001). Revealed preference tech-iques estimate the approximate expenditures that are involved
n engaging in activities like tourism. Revealed preference tech-iques include travel cost and hedonic pricing, as well as, variousechniques for substituting expenditures related to remediation orroduction improvement of the service in question (Swinton et al.,007). Travel cost estimates determine the value of a location-pecific service by assuming that travel expenditures influenceemand (Hein, 2011; Martín-López et al., 2009). Hedonic valuation
s a technique where land price are compared to location-specificharacteristics. Difference in land prices are used to approximatehe value of services like scenic views and proximity to waterCampbell, 2007; Cavailhès et al., 2009; Swinton et al., 2007).
Of the available monetary valuation techniques, we have cho-en a stated contingent procedure. The method was considered toe suitable for the sample group (less time consuming and cogni-ively challenging) as tourist are often more interested in leisureime than taking part in a questionnaire. In addition to this statedreference estimate we calculate the revealed preference by using aime/cost estimate, asking where respondents have travelled from,o approximate the actual expenditures of travelling to the regiono enjoy the cultural services.
.5.2. Stated preference value estimateThe WTP exercise was developed to allow respondents to com-
are possible landscape changes for assessing the importance andorth of landscape maintenance sustaining cultural ecosystem
ervices. Panoramic photos of existing landscapes including theraditional agricultural, coulissen landscape and extensive graz-ng lands were altered using Adobe PhotoshopTM. This ensuredhat weather conditions and ambient light were constant betweenhe photos, preventing that these extraneous factors play a rolen choice preferences. Three important processes that will likelyhange the landscape character in the future are assessed: (a)ncreased residential infill with increasing urban in-migration; (b)he cutting of tree lines and hedgerows for scale-enlargement ingricultural production; and (c) rewilding due to agricultural aban-onment (Fig. 4). New features like meadow grasses, housing andractors were introduced as novel landscape elements representinghe changes while in some photos tree lines and hedgerows wereaken out.
In addition to the visual comparison, respondents were alsoiven an explanation of the processes leading to these landscapehanges. This was to establish a payment vehicle by which respon-ents understood how their contribution would contribute to
andscape conservation (e.g. by clarifying that a yearly contribu-ion from their taxes would go to farmers as a cost for maintaininghe landscape). It was explained that many farmers were experi-ncing financial difficulty due to the current economic climate andotentially reduced subsidies. As part of farm survival strategies,ome were presumed to resort to agricultural intensification whilethers were presumed to stop farming all together. These farmerecisions were linked to the landscape changes depicted in the pho-os. A landscape maintenance subsidy was proposed as a way tougment farmers’ income and maintain the current landscape aes-hetics. Respondents were asked how much they would be willingo contribute each year to such a landscape maintenance fund.
.5.3. Revealed preference value estimateThe estimate of revealed preference was made by calculating
time cost demand curve for the sample group. This approach
al Indicators 37 (2014) 163– 174
assumes that there is decreasing demand for visiting the regiondue to travel cost. A similar pricing procedure has been employedin other revealed preference valuation studies (Hein, 2011; Martín-López et al., 2009). Respondents’ travel costs are calculatedaccording to a 34 eurocent/km rate, which has been used in othervaluation studies in the Dutch context (Hein, 2011). An averagehourly wage of 14.70 euro is applied for travel time based on theaverage income of the sample and a 40 h work week. Travel timeis calculated according to estimates in Google Maps. It is assumedthat visitors travel by car to the area. Out of the sample group, themajority of respondents travelled by car (84%) while some respon-dents also travelled by public transport and bicycle (recreants). Thedemand curve is modelled based on the proportion of respondentsvisiting the region from different travel cost zones (Table 3). Thetotal population of the different distance zones is divided by actualvisits. The distribution of demand of the sample is assumed to rep-resent total annual visits of all visitors to camping site and bed andbreakfast as obtained from municipal records. From this visitationcurve an estimate of the cost function can be extrapolated assumingthat current visitation represents total demand. The demand curvewas calculated using a simple linear extrapolation of increasedtravel cost. Based on this demand curve extrapolation, consumersurplus could be estimated by calculating the area under the curve.
3. Results
3.1. Respondent characteristics and preferences
Cycling and walking are the main activities that attract touristsand recreants to the case study area (Table 1). Respondents oftenqualified this expressing that the attractiveness of the landscapeenhanced their enjoyment of such activities. Tranquility and restwas also an activity mentioned by a number of respondents. Tran-quility scored high despite not being an option from the list ofactivities on the questionnaire. Swimming, shopping and eatingwere also often chosen and indicated as side activities to biking andwalking. The results of the photo assessment of preferred landscapefeatures are shown in Fig. 2. Cultural buildings and tree lines wererated high and were often immediately recognised as distinctive forthe region. Respondents also appreciated ‘natural’ features includ-ing brooks, forests and the knowledge that wild animals lived inthe area. Respondents assessment of landscape structure revealedthat forest patches interconnected with tree lines was most pre-ferred (n = 56). A similar landscape structure with slightly less forestpatches was second most preferred (n = 26). Those seeking leisurein the region were not overly interested in forest dominated land-scape (n = 11) indicating a preference for a natural mosaic landscape(Fig. 3).
3.2. Cultural service maps
3.2.1. Preference hot spotsThe map of preference for landscape features indicates a num-
ber of hot spots where numerous features that are preferredby respondents are located (Fig. 5). The municipality of Winter-swijk and borders of Berkellend and Oost Gelre are areas withhigh values. This is due to the coincidence of numerous land-scape features including tree lines, forests, cultural buildings andanimal habitats in these locations. In the centre of map valuesare lower. These cold spots can be characterized as locationswhere there is an absence of visible animal habitat and where
the landscape is dominated by open agriculture land and modernlarge scale farm businesses. In the map depicting preference forlandscape structure a similar pattern is apparent. The similarityindicates that there is substantial overlap between the locations of
ologic
ltafiftds
Fa
D.B. van Berkel, P.H. Verburg / Ec
andscape features and the landscape structure and compositionhat are valued by tourist respondents. Notable exceptions arereas that are dominated by forests. While in the features map,orests are indicated with moderately high values, the same areasn the structure map have low values. Respondents appreciate
orest as landscape features but in terms of landscape structurehey prefer mosaic landscapes with smaller areas of forest. Suchifference cannot be distinguished in the analysis based on land-cape features alone as the coincidence of forest and wildlifeig. 5. Maps depicting preference for landscape (a), features (b) and structure and composrea.
al Indicators 37 (2014) 163– 174 169
habitat contributes to higher scores in the landscape featuremap.
3.2.2. Respondent groups and cultural servicesRespondents’ assessment of the landscape function showed that
aesthetic beauty (x̄ = 4.70) and recreation (x̄ = 4.16) are highlyvalued services provided by the landscape in the region. Cul-tural heritage (x̄ = 3.70), inspiration (x̄ = 3.27) and spirituality(x̄ = 2.38) where rated less important. Statistical exploration of all
ition as indicator of the values of cultural service; and (c) land use of the case study
170 D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174
Table 1Respondents assessment of tourist attractions in the Achterhoek region.
Top activity Second activity Third activity Weighted value
Cycling 68 18 8 248Walking 12 41 7 125Swim 13 11 4 65Tranquility and rest 9 5 8 45Shopping 0 9 14 32Eat and drink 0 4 20 28Farm-based camping 2 4 6 20Unique landscape 1 3 6 15Visit family 3 0 5 14Region specific recreational activities 2 2 1 11Festival 1 1 4 9Other 5 4 5Nothing 0 4 18
N m to ta
rai4t(nintca(4t
gmph(AtfTtccur
tfs
TP
ote – Respondents were required to pick the top three activities that attracted thend 1 to the second and third answers respectively.
esponses revealed a number of determining factors for landscapeppreciation. An independent t-test indicated that respondents liv-ng within 30 min of the case study area valued recreation (x̄ =.59), spirituality (x̄ = 2.83) and inspiration (x̄ = 3.72) higher thanhose living further away (x̄ = 4.01; x̄ = 2.33; x̄ = 3.12 respectively)t = 3.54, p < 0.001, n = 81; t = 1.93, p < 0.1, n = 113; t = 2.59, p < 0.01,
= 58). For Dutch respondents (x̄ = 3.83) cultural heritage is moremportant in comparison to Germans (x̄ = 2.60) (t = 3.33, p < 0.01,
= 18). Similarly those visiting agri-camping sites appreciate cul-ural heritage (x̄ = 4.50) more than those staying in traditionalamping sites (x̄ = 3.45) (t = 3.26, p < 0.01, n = 80). Age also plays
role as respondents older than 50 appreciate the aesthetic beautyx̄ = 4.83) of the landscape more than younger respondents (x̄ =.53) (t = 3.10, p < 0.01, n = 113). Repeat visitors to the area (>10imes) tended to have a higher appreciation of the cultural services.
The exploration of responses by way of a PCA uncovered threeroups of like respondent preferences using the quartimax rotationethod. Three components were retained despite the third com-
onent’s low eigenvalue (0.85) as commonalities were adequatelyigh for all variables (<0.7). A total KMO revealed a mediocre fit0.65); however, the Bartlett’s test result was significant (<0.001).s the PCA was primarily used to group like clusters of respondents
hese results were deemed acceptable. A strong component scoreor each component also demonstrated useful groupings (Table 2).he first component is a group of respondents who rate inspira-ion and spiritual highly (total variance (tv) 30.4%). The secondomponent comprised respondents that score aesthetic beauty andultural heritage highly (tv = 29.4%). The third component is madep of respondents who highly appreciate the landscape as a site forecreation (tv = 20.65).
The ANOVA t-test on the component groups revealed that
here was substantial variation regarding preferences for landscapeeatures (supplementary material). Respondents appreciating thepiritual and inspirational qualities of the landscape rated tree linesable 2rinciple component analysis of landscape function appreciation.
Landscape function Component
1 2 3
Recreation 0.064 0.000 0.983Aesthetic beauty 0.204 0.825 0.128Cultural heritage 0.147 0.848 −0.122Inspiration 0.847 0.221 −0.096Spirituality 0.857 0.140 0.159Eigenvalues
Total 2.103 1.066 0.852% of variance 42.061 21.329 17.037Cumulative % 42.061 63.390 80.427
he region. The weight value is calculated by applying a score of 3 to top answers, 2
significantly higher (x̄ = 1.40) than all other respondents (x̄ = 0.83)(t = 2.35, p < 0.05, n = 76). Those with a strong appreciation of theaesthetic beauty of the landscape and cultural heritage value forest(x̄ = 1.08) and animals (x̄ = 0.92) more than the other groups (x̄ =0.55; x̄ = 0.47) (t = 2.47, p < 0.05, n = 113; t = 2.32, p < 0.05, n = 113).They have no regard for recreation facilities in the landscape (x̄ = 0)comparing all others (x̄ = 0.71) (t = −2.14, p < 0.05, n = 113). Respon-dents valuing the recreation possibilities provided by the landscapefind an agricultural landscape less important (x̄ = 0.30) in compari-son to the other groups (x̄ = 0.61) (t = −1.70, p < 0.10, n = 113). Therewas no difference between the groups concerning preference forlandscape structure and composition.
3.2.3. Stated valueThe mean WTP per year for landscape maintenance in Winter-
swijk is D 86.18 per person (std. dev. 125.87) removing extremeoutliers (1 respondent with a WTP more than 7% of his/her totalnet income was removed). Respondents most valued the conserva-tion of the coulissen landscape. They were on average WTP D 33.30(std. dev. 83.37) to prevent farmers from cutting landscape ele-ments to improve agricultural productivity. The conservation ofthe other landscapes was valued slightly less. Respondents wouldcontribute D 27.30 (std. dev. 59.56) to conserve traditional agricul-ture landscapes from increased residential infill and D 23.87 (std.dev. 57.01) to prevent that extensive farming landscapes becomeovergrown and wild. Low WTP for the conservation of extensivefarmland is attributed to the fact that respondents did not find therewilding scenario problematic. When asked to rank the attractive-ness of the landscape changes depicted in photo-manipulations,rewilding was rated highest by 62% of the respondents (Fig. 6).Increased urbanisation at the expense of the traditional landscapewas viewed as least attractive by 57% of respondents. The picturedepicting a landscape with removed tree lines for increased agri-cultural productivity was ranked in the middle (60% of responses).This result contradicts respondents WTP for preserving the coulis-sen landscape, which received the largest monetary value. Of thetotal sample, 50% (n = 57) of the respondents were not willing to payfor the maintenance of the landscape. Many of those respondentswere protest bidders who stated that they valued the landscapebut were unwilling to contribute to its conservation. Respondentscited different reasons for their protest bid including a mistrustof government spending, low income, a conviction that currentbudget could cover landscape incentives and the uncertainty thatinvestment would yield described services.
3.2.4. Revealed valueBased on the total number of visit per year presented in Table 3
and the assumption that our respondent sample is representative
D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174 171
Table 3Visitor rates and travel costs to Winterswijk.
Zone Percent ofrespondents
Estimated total visit/yearbased camping and B&B visits
Zone population Annual visit/1000people
Average travelcosts (D )
30 km 16 5907 243 772 24.23 6.7030–60 km 16 5907 1 133 989 5.21 30.4460–100 km 23 8532 7 660 047 1.11 46.43100–150 km 13 4594 12 275 825 0.37 66.67150–200 km 23 8532 13 219 904 0.65 87.21200–250 km 9 3282 14 814 175 0.22 111.45
Total 100 36 755
0
10
20
30
40
50
60
70
80
Mostpreferred
Middlepreference
Lea stpreferred
Fequ
ency Residental infill
Removal of landscapeelements
Rewilding
oart
V
fvaoievv
Fig. 6. Respondents’ preference for landscape evolution processes.
f total annual visitors trips of campers and those staying in bednd breakfast (n = 36,755), the relation between travel cost and visitate is calculated. Eq. (1) describes the visit rate as a function of theravel cost for the municipality of Winterswijk.
isit rate = 16.804e−0.043∗cost (R2 = 0.85) (1)
The demand function is calculated assuming that expendituresor travel are viewed as the cost to experience/partake in the ser-ices located in the study area. This does not include expendituresssociated with lodging, which can be considered part of total costf staying in the area to enjoy the cultural services. The result-ng demand curve is presented in Fig. 7. The area under the curvequalling the lower bound consumer surplus of the landscape ser-ices is approximately D 850 000/year. This equals around D 23 perisit.
y = -23.2ln(x) + 244.4
-20
0
20
40
60
80
100
120
140
160
4000035000300002500020000150001000050000
Add
ed c
osts
/tri
p (e
uro)
Annual visits
Fig. 7. Demand curve for tourists visiting Winterswijk.
49 347 712
4. Discussion
4.1. Overview
In this paper we have demonstrated a method for mapping andquantifying the provision of cultural services at the regional scale.By collecting indicators that could be linked to cultural serviceswe were able to show spatial hotspots of cultural service deliveryand make a lower boundary monetary estimate of services origi-nating from the landscape. These different valuation techniques arerarely combined (Naidoo and Ricketts, 2006; Willemen et al., 2010),despite recognition that spatial monetary valuation is importantfor effective land management (Daily et al., 2009). Often, stud-ies include spatial monetary values of services based on simpleper hectare estimates derived from meta-analysis of case studiesand/or national/global approximations (Lautenbach et al., 2011;Naidoo and Ricketts, 2006; Willemen et al., 2010). Per hectare val-ues are then translated to pixel or administrative units. In our studywe differentiate how landscape features and the structure of thelandscape are valued. Based on these preferences we are able todiscriminate between highly valued landscapes and less valuedlandscapes as well as, being able to understand the influence of indi-vidual landscape elements and their spatial structure. Contingentmonetary valuation was also based on location specific charac-teristics. This spatial specificity of regional assets is important forbalancing planning initiatives by allowing for the specification ofsuitable areas for developments and tailoring tourist amenities tospecific groups.
The results show that regions that retain landscape featureslike cultural buildings, tree lines, lakes and rivers, forests andwildlife viewing are appreciated by visitors. Semi-managed land-scape structures composing of forest patches interlinked withhedgerows have likewise been identified as important. The mon-etary estimate of these services, based on travel cost, place thelower bound value at D 850 000/year for the municipality of Win-terswijk alone. An estimate for the conservation of the landscapebased on respondents WTP for maintaining current landscapes isD 3.2 million. Given respondent bias such as warm glow (i.e. overes-timation of WTP based on social stigma), the actual value of servicemight be more realistically placed somewhere in between thesefigures. It should be noted that these estimates only include theassessed cultural services. The landscape provides multiple func-tions in addition to cultural services. A total expenditure estimatefor the income produced from tourism is probably much higher. Inthe study we do not account for lodging and other tourist expendi-tures such as those in the restaurant and entertainment industries.
To give an estimate of the cultural services for the entire casestudy region we extrapolate landscape preferences from Win-terswijk to the other municipalities. We assume that landscape
preference is related to the monetary values derived for the WTPexercise, as wells the travel cost estimate. Table 4 gives the perhectare value of the landscape distributing the total estimatedvalue of cultural service as proportion of the feature valuation
172 D.B. van Berkel, P.H. Verburg / Ecological Indicators 37 (2014) 163– 174
Table 4Total and per hectare estimate of the monetary value of ES per municipality.
Estimate of monetary value ofES based on WTP (total)
Estimate of monetary value ofES based on travel cost (total)
Estimate of per hectare monetaryvalue of ES based on WTP
Estimate of per hectare monetary valueof ES based on travel cost
Winterswijk 850 000 3.2 million 0.62D /h 2.31D /hAalten 490 000 1.8 million 0.50D /h 1.90D /h
mtahaufs
tsStisd
4
ta2vriada1so
Oot Gelre 510 000 1.9 million
Berkelland 1.1 million 4.2 million
Total 2.95 million 11.1 million
ap. A comparison of the monetary value of cultural service forhe different municipalities indicates that there is heterogeneity inssets. The municipality of Winterswijk has the highest value perectare due to the abundance of appreciated landscape featuresnd composition. Aalten, Oost Gelre and Berkelland are less val-ed. Such empirical evidence indicates that proposed incentivesor landscape maintenance in Winterswijk are justified if they areuccessful in preserving these cultural services.
Most ecological indicators are based on ecological characteris-ics (i.e. species numbers, landcover), while not often consideringocietal preference for the associated services of these ecosystems.ocietal preferences can dictate ecological processes in agricul-ural landscapes through human management of ecosystems. Thentegration of societal preferences into ecological indicators is atep toward providing policymakers insights into these ecologicalrivers (MA, 2003).
.2. Quantifying cultural services
Many studies stress the importance of the quantification of cul-ural services while actual valuation is often limited to tourismmenities (Egoh et al., 2008; Nelson et al., 2009; Willemen et al.,008). This is due to the difficulty in estimating how respondentsalue intrinsic characteristics like inspiration. Our investigationevealed that cultural heritage, aesthetic beauty, spirituality andnspiration play a role in attracting different tourists and recre-nts. Economic valuation studies have not been able to address thisifferentiated valuation of intrinsic qualities. Monetary estimates
re usually given to broader environmental (Brouwer and Slangen,998) and tourist services (Martín-López et al., 2009) where intrin-ic qualities are assumed to be valued and included. The dangerf ignoring these differences is that characteristics associated withFig. 8. Comparison of different group preference for landscape features
0.46D /h 1.74D /h0.42D /h 1.63D /h0.50D /h 1.90D /h
intrinsic qualities (cultural buildings, tree lines) are not valued forthe important contribution that they make to total services deliv-ery.
While our findings show that groups appreciate intrinsic qual-ities differently, only a weak link could be made to landscapefeatures appreciated by them (supplementary material). Groupscould not be differentiated according to their appreciation of thestructure and composition of the landscape as indicated by theirmapped preferences (Fig. 8). This was also the case in the evalu-ation of the appreciation of the different trajectories of landscapeevolution (Fig. 7). High homogeneity in preferences and aversion tocertain developments is likely the cause of this result. It may wellbe that when more subtle differences between landscape evolu-tion would be evaluated there would be more difference betweenthe groups. Howley et al. (2012) for instance found that a num-ber of different socio-economic factors contribute to difference inthe preference for small differences in landscape structure. Thisraises the question: can intrinsic qualities be usefully parame-terised with spatial proxies? A number of spatial studies haveaddressed intrinsic service localisation (Alessa et al., 2008; Brownand Raymond, 2007; Sherrouse et al., 2011). Findings by Alessa et al.(2008) show that local community members demonstrate a highlydeveloped sense of place for which they can differentiate loca-tions with high capacity for intrinsic qualities like spirituality andinspiration. In the same study there was variation between commu-nities in locating these areas, suggesting a highly spatial componentin respondents’ assessments where awareness and familiarity areimportant (Soini et al., 2012). In our study we aimed at defining
more generic features using landscape photos, where spatial famil-iarity is not necessary. In contrast to residents, tourists are oftenmuch less familiar with maps of the region and are therefore lesscapable to indicate on maps locations with high value. Our method(top row) and landscape structure and composition (bottom row).
ologic
daatbabal
tp2mramalefdpllreots
4p
trPiia2etRs2ipfceTbifgwi
boeia2
D.B. van Berkel, P.H. Verburg / Ec
emonstrates a first step in unravelling how landscape functionppreciation can be related to landscape features and structurend composition (see also Howley et al., 2012) and their loca-ion (Dramstad et al., 2006). The results also clearly illustrate thatoth individual features and the overall structure of the landscapere important and address different aspects of appreciation. Mapsased on either of these approaches showed therefore, in somereas, contradicting results, providing more insight in the actualandscape characteristics and determinants of landscape value.
A drawback of assessing landscape value based on spatial cri-eria is that this does not consider rural dynamics where differentrocesses and actors influence change in the landscape (Wilson,010; Van Berkel et al., 2011). It likewise does not provide infor-ation regarding the value placed on conserving a landscape in
elation to the policy input needed. The WTP assessment givesn indication of the trade-off that societal actors are willing toake for the conservation of certain landscapes. Landscape man-
gers and planners are often confronted with the dilemma thatocal budgets take precedence over maximisation of societal ben-fit. Understanding which landscapes are more valued is usefulor targeting resources efficiently. In the Dutch case study, rewil-ing was not seen as overly problematic and this is supported byreference for semi-managed landscape structure. Recent findings
ikewise suggest that medium and high levels of succession withess human intervention are appreciated (Howley et al., 2012). Theesults of landscape composition and structure appreciation how-ver indicate that large scale rewilding, leading to larger patchesf natural vegetation may not be preferred. Such results indicatehat while preventing re-wilding may not need policy priority, largecale forest regrowth will need attention.
.3. The use of photo and photo-realistic montage as respondentrompts
It was hypothesised that photos would give an added dimensiono questionnaire inquiry by helping in eliciting honest responsesegarding respondent’s ideas and feelings about the landscape.hoto and photo manipulated images are increasingly employedn participatory planning for helping in creating stakeholder buy-n (Lovett et al., 2010; Soliva et al., 2010; Van Berkel et al., 2011),nd to illicit preference for landscape aesthetics (Dramstad et al.,006). In this study they were effective. Photo of features wereasily recognisable for respondents and they often commentedhat they enjoyed the exercise of ranking and comparing images.epresenting landscape structure through aerial photographs wasimilarly accepted by respondents (see also Fagerholm and Käyhkö,009), despite reservation that the aerial view would be confus-
ng for those unfamiliar with such spatial representations. Theopularity of Google Earth/Maps may account for this. We alsoound that the landscape photos depicting the landscape evolutiononveyed rich meanings that respondents could decipher. This wasvident in respondent comparison of current and predicted photos.hey often made specific comments about the features that hadeen added or taken away addressing issues such as the density of
ntroduced housing, the extent of rewilding and the perspective ofarmers for landscape management. The findings of this study sug-est that using photos of ecological characteristics is an effectiveay to integrate societal preferences for landscape characteristic
nto ecological indicators.The reliability of the accuracy of obtaining aesthetic preference
y way of photos could not be judged, but findings are similar tother studies assessing visual preference for landscapes (Howley
t al., 2012; Ode et al., 2009). Bias in responses is often an issuen preference and valuation studies. Weather condition and relax-tion influence respondents’ answers (De Groot and van den Born,003). In our assessment with tourists this bias may have been anal Indicators 37 (2014) 163– 174 173
issue. However, respondents did acknowledge that they appreciatelocal assets more in good weather conditions, suggesting that theirresponses took this into account.
5. Conclusions
The future of cultural landscapes is uncertain as both endoge-nous and exogenous processes will play a role in their futurefunctionality. This study demonstrates that there is societaldemand for the cultural services that such agricultural landscapesprovide. Their continued resilience will require understanding thedemands for service so that processes that might hinder their pro-vision can be intervened upon. Spatial understanding of the assetsdelivered by the landscape can help in prioritizing areas for mainte-nance/restoration strategies while demonstrating the importanceof conservation of cultural service delivery.
Acknowledgements
This research has been conducted in the context of the DutchKnowledge for climate (KvK) PROJECT and the EU funded FP7 CLAIMproject. It is a contribution to the Global Land Project. The Authorswould like to thank Anouk Adang and Eleni Dellas for their help inconducting the survey. We would also like to thank all those whotook part in the questionnaire.
Appendix A. Supplementary data
Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.ecolind.2012.06.025.
References
Alessa, L., Kliskey, A., Brown, G., 2008. Social–ecological hotspots mapping: a spatialapproach for identifying coupled social–ecological space. Landscape Urban Plan.85, 27–39.
Birol, E., Koundouri, P., Kountouris, Y., 2008. Integrating wetland management intosustainable water resources allocation: the case of Akrotiri wetland in Cyprus.J. Environ. Plan. Manage. 51, 37–53.
Brouwer, R., Slangen, L.H.G., 1998. Contingent valuation of the public benefits ofagricultural wildlife management: the case of Dutch peat meadow land. Eur.Rev. Agric. Econ. 25, 53–72.
Brown, G., 2006. Mapping landscape values and development preferences: a methodfor tourism and residential development planning. Int. J. Tourism Res. 8,101–113.
Brown, G., Raymond, C., 2007. The relationship between place attachment and land-scape values: toward mapping place attachment. Appl. Geogr. 27, 89–111.
Bryan, B., Grandgirard, A.S., Ward, J., 2010. Quantifying and exploring strategicregional priorities for managing natural capital and ecosystem services givenmultiple stakeholder perspectives. Ecosystems 13, 539–555.
Buijs, A., Pedroli, B., Luginbühl, Y., 2006. From hiking through farmland to farmingin a leisure landscape: changing social perceptions of the European landscape.Landscape Ecol. 21, 375–389.
Campbell, D., 2007. Willingness to pay for rural landscape improvements: combiningmixed logit and random-effects models. J. Agric. Econ. 58, 467–483.
Cavailhès, J., Brossard, T., Foltête, J.-C., Hilal, M., Joly, D., Tourneux, F.P., et al., 2009.GIS-based hedonic pricing of landscape. Environ. Resour. Econ. 44, 571–590.
Colombo, S., Calatrava-Requena, J., Hanley, N., 2006. Analysing the social benefitsof soil conservation measures using stated preference methods. Ecol. Econ. 58,850–861.
Daily, G.C., Polasky, S., Goldstein, J., Kareiva, P.M., Mooney, H.A., Pejchar, L., et al.,2009. Ecosystem services in decision making: time to deliver. Front. Ecol. Envi-ron. 7, 21–28.
De Groot, W.T., van den Born, R.J.G., 2003. Visions of nature and landscape type pref-erences: an exploration in The Netherlands. Landscape Urban Plan. 63, 127–138.
Dockerty, T., Lovett, A., Nnenberg, G., Appleton, K., Parry, M., 2005. Visualising thepotential impacts of climate change on rural landscapes. Comput. Environ. UrbanSyst. 29, 297–320.
Dramstad, W.E., Tveit, M.S., Fjellstad, W.J., Fry, G.L.A., 2006. Relationships between
visual landscape preferences and map-based indicators of landscape structure.Landscape Urban Plan. 78, 465–474.Egoh, B., Reyers, B., Rouget, M., Richardson, D.M., Le Maitre, D.C., van Jaarsveld,A.S., 2008. Mapping ecosystem services for planning and management. Agric.Ecosyst. Environ. 127, 135–140.
1 ologic
F
G
H
H
H
I
K
K
L
L
M
M
M
M
M
M
N
N
N
O
O
P
P
ization of landscape functions. Landscape Urban Plan. 88, 34–43.
74 D.B. van Berkel, P.H. Verburg / Ec
agerholm, N., Käyhkö, N., 2009. Participatory mapping and geographical patterns ofthe social landscape values of rural communities in Zanzibar, Tanzania. Fennia187, 46–60.
eoghegan, J., Wainger, L.A., Bockstael, N.E., 1997. Spatial landscape indices in ahedonic framework: an ecological economics analysis using GIS. Ecol. Econ. 23,251–264.
anley, N., Mourato, S., Wright, R.E., 2001. Choice modelling approaches: a superioralternative for environmental valuation? J. Econ. Surv. 15, 435–462.
ein, L., 2011. Economic benefits generated by protected areas: the case of the HogeVeluwe Forest, The Netherlands. Ecol. Soc. 16 (2):13.
owley, P., Donoghue, C.O., Hynes, S., 2012. Exploring public preferences for tradi-tional farming landscapes. Landscape Urban Plan. 104, 66–74.
EEP (Institute for European Environmental Policy), 2007. Final Report on the Studyof HNV Indicators for Evaluation. European Commission, DG Agriculture, Brus-sels.
inzig, A.P., Perrings, C., Chapin, F.S., Polasky, S., Smith, V.K., Tilman, D., et al., 2011.Paying for ecosystem services? Promise and peril. Science 334, 603–604.
nickel, K., Renting, H., Van der Ploeg, J.D., 2004. Multifunctionality in Europeanagriculture. In: Brouwer, F. (Ed.), Sustaining Agriculture and the Rural Envi-ronment, Governance, Policy and Multifunctionality. Edward Elgar Publishing,Cheltenham, UK, pp. 81–103.
autenbach, S., Kugel, C., Lausch, A., Seppelt, R., 2011. Analysis of historic changesin regional ecosystem service provisioning using land use data. Ecol. Indic. 11,676–687.
ovett, A.A., Ribeiro, S.C., Van Berkel, D.B., Verburg, P.H., Firmino, A., 2010. Repre-senting and communicating rural futures through 3D landscape visualizations:experiences from the RUFUS project. In: Pietsch, M., Kretzler, E. (Eds.), PeerReviewed Proceedings of Digital Landscape Architecture 2010. Anhalt Universityof Applied Sciences, Wichmann Verlag, Heidelberg, pp. 261–268.
A, 2003. Millennium Ecosystem Assessment, Ecosystems and Human Well-being:A Framework for Assessment. Island Press, Washington.
a, S., Swinton, S.M., 2011. Valuation of ecosystem services from rural landscapesusing agricultural land prices. Ecol. Econ. 70, 1649–1659.
ari Sundli, T., 2009. Indicators of visual scale as predictors of landscape preference;a comparison between groups. J. Environ. Manage. 90, 2882–2888.
arsden, T., 1999. Rural futures: the consumption countryside and its regulation.Sociol. Ruralis 39, 501–526.
arsden, T., Sonnino, R., 2008. Rural development and the regional state: denyingmultifunctional agriculture in the UK. J. Rural Stud. 24, 422–431.
artín-López, B., Gómez-Baggethun, E., Lomas, P.L., Montes, C., 2009. Effects of spa-tial and temporal scales on cultural services valuation. J. Environ. Manage. 90,1050–1059.
aidoo, R., Ricketts, T.H., 2006. Mapping the economic costs and benefits of conser-vation. PLoS Biol. 4.
edkov, S., Burkhard, B., 2011. Flood regulating ecosystem services – mapping sup-ply and demand, in the metropole municipality, Bulgaria. Ecol. Indic. 21, 67–79.
elson, E., Mendoza, G., Regetz, J., Polasky, S., Tallis, H., Cameron, D.R., et al., 2009.Modeling multiple ecosystem services, biodiversity conservation, commodityproduction, and tradeoffs at landscape scales. Front. Ecol. Environ. 7, 4–11.
’Farrell, P.J., Anderson, P.M., 2010. Sustainable multifunctional landscapes: areview to implementation. Curr. Opin. Environ. Sustainability 2, 59–65.
de, Å., Fry, G., Tveit, M.S., Messager, P., Miller, D., 2009. Indicators of perceivednaturalness as drivers of landscape preference. J. Environ. Manage. 90, 375–383.
hilip, D., 1984. Factors influencing landscape preferences: an empirical investiga-tion. Landscape Plan. 11, 293–306.
into-Correia, T., Gustavsson, R., Pirnat, J., 2006. Bridging the gap between centrallydefined policies and local decisions – towards more sensitive and creative rurallandscape management. Landscape Ecol. 21, 333–346.
al Indicators 37 (2014) 163– 174
Provincie Gelderland, 2010. Geografische gegeven (Geographical data). PovincieGelderland, Arnhem.
Raymond, C.M., Bryan, B.A., MacDonald, D.H., Cast, A., Strathearn, S., Grandgirard,A., et al., 2009. Mapping community values for natural capital and ecosystemservices. Ecol. Econ. 68, 1301–1315.
Ready, R.C., Berger, M.C., Blomquist, G.C., 1997. Measuring amenity benefitsfrom farmland: hedonic pricing vs. contingent valuation. Growth Change 28,438–458.
Renting, H., Rossing, W.A.H., Groot, J.C.J., Van der Ploeg, J.D., Laurent, C., Perraud,D., et al., 2009. Exploring multifunctional agriculture. A review of conceptualapproaches and prospects for an integrative transitional framework. J. Environ.Manage. 90, S112–S123.
Ryan, R.L., 2011. The social landscape of planning: integrating social and percep-tual research with spatial planning information. Landscape Urban Plan. 100,361–363.
Santana-Jimènez, Y., Suárez-Vega, R., Hernàndez, J.M., 2011. Spatial and environ-mental characteristics of rural tourism lodging units. Anatolia 22, 89–101.
Sherrouse, B.C., Clement, J.M., Semmens, D.J., 2011. A GIS application for assessing,mapping, and quantifying the social values of ecosystem services. Appl. Geogr.31, 748–760.
Soini, K., Vaarala, H., Pouta, E., 2012. Residents’ sense of place and landscape per-ceptions at the rural–urban interface. Landscape Urban Plan. 104, 124–134.
Soliva, R., Bolliger, J., Hunziker, M., 2010. Differences in preferences towards poten-tial future landscapes in the Swiss Alps. Landscape Res. 35, 671–696.
Soliva, R., Rønningen, K., Bella, I., Bezak, P., Cooper, T., Flø, B.E., et al., 2008. Envision-ing upland futures: stakeholder responses to scenarios for Europe’s mountainlandscapes. J. Rural Stud. 24, 56–71.
Solymosi, K., 2011. Indicators for the identification of cultural landscape hotspots inEurope. Landscape Res. 36, 3–18.
Stenseke, M., 2009. Local participation in cultural landscape maintenance: lessonsfrom Sweden. Land Use Policy 26, 214–223.
Swinton, S.M., Lupi, F., Robertson, G.P., Hamilton, S.K., 2007. Ecosystem services andagriculture: cultivating agricultural ecosystems for diverse benefits. Ecol. Econ.64, 245–252.
Tress, B., Tress, G., 2003. Scenario visualisation for participatory landscape planning– a study from Denmark. Landscape Urban Plan. 64, 161–178.
Van Berkel, D.B., Carvalho-Ribeiro, S., Verburg, P.H., Lovett, A., 2011. Identify-ing assets and constraints for rural development with qualitative scenarios:a case study of Castro Laboreiro, Portugal. Landscape Urban Plan. 102,127–141.
Van Berkel, D.B., Verburg, P.H., 2011. Sensitising rural policy: assessing spa-tial variation in rural development options for Europe. Land Use Policy 28,447–459.
Van den Berg, A.E., Koole, S.L., 2006. New wilderness in the Netherlands: an inves-tigation of visual preferences for nature development landscapes. LandscapeUrban Plan. 78, 362–372.
Wildenbeest, G., 1989. Recent farmers’ protest in a Dutch municipality: the legacyof the past. In: Boissevain, J., Verrips, J. (Eds.), Dutch Dilemmas: Anthropologistslook at the Netherlands. Van Gorcum, Assen, pp. 70–88.
Willemen, L., Hein, L., Verburg, P.H., 2010. Evaluating the impact of regional devel-opment policies on future landscape services. Ecol. Econ. 69, 2244–2254.
Willemen, L., Verburg, P.H., Hein, L., van Mensvoort, M.E.F., 2008. Spatial character-
Wilson, G., 2010. Multifunctional ‘quality’ and rural community resilience. Trans.Inst. Br. Geogr. 35, 364–381.
Zimmermann, R.C., 2006. Recording rural landscapes and their cultural associations:some initial results and impressions. Environ. Sci. Policy 9, 360–369.
Recommended
