1. Analyzing and mapping cultural ecosystem services with multiple approaches in Peru
Session T4: Mapping ecosystem services: Comparing methods. October 21, 2016
Bruno Locatelli1,2; Merelyn Valdivia2,3; Améline Vallet1,2,4
1: CIRAD France, 2: CIFOR Peru, 3: UNALM Peru, 4: AgroParisTech France
Bruno Locatelli1,2; Raffaele Vignola3; Diego Padilla2
1: CIRAD France, 2: CIFOR Peru, 3: CATIE, Costa Rica

2. Modeling and mapping multiple ecosystem services
Provisioning
Regulation
Cultural
Problems for modeling and mapping:
Limited data; Multidimensional (what values?); Subjective (whose values?).

3. The spirit of the mountain
Cultural services
I want to climb
Recreation
(experiential)
What a nice view!
Scenic beauty
(visual)
The spirit of the mountain
Others
Heritage and legacy.
Spiritual and religious.
Inspiration and creativity.
Identity and sense of place.
Social relationships.
Education and knowledge.
(intellectual)

4. Case study: Peru Questions:
What cultural ecosystem services are produced and used? Where? Why there?
Who use them?
Case study:
Peru
Apurímac

5. Multiple methods: Each one provides a piece of the puzzle
Big data
and maths
Local fieldwork and social science

6. Density of online landscape picture databases
Big data
and maths
Density of online landscape picture databases
Method:
Automatically download online pictures georeferenced in the region
Already filtered by Panoramio for Google Earth
No pizzas or parties
Mostly landscapes (natural, rural, urban)
Analyze and map density

7. Good representation of where ecosystem services are “used”
Results
Good representation of where ecosystem services are “used”
Do it for all GE pics: Add:
** ROAD
C:\Users\BLocatelli\Documents\Bruno\a-Metadata\r. Peru\Peru Red de Vias\RVN\RVN_Eje.shp
Geometry Type: Line
COD_VIAL = PE-3S
** CITIES
C:\Users\BLocatelli\Documents\Bruno\a-Current\Res-Abancay\Data_Maps\ZEE maps\B_CCPP.shp
Geometry Type: Point
V_CAT_PUEB = Capital*
** NATIONAL PARK
C:\Users\BLocatelli\Documents\Bruno\a-Metadata\r. Peru\Peru Areas Protegidas SERNANP\anp_nacional_05_08_2016\anp_nacional_05_08_2016.shp
Geometry Type: Polygon
FID=1
** RIVER
In ArcGIS: Major rivers in allRios (ESTADO ==2)
MERGE WITH B_RED_HID select =
"FID" =2283 OR "FID" =2300 OR "FID" =2945 OR "FID" =10937 OR "FID" =10996 OR "FID" =11066 OR "FID" =11147 OR "FID" =11326 OR "FID" =11445 OR "FID" =11458 OR "FID" =11643 OR "FID" =11959 OR "FID" =12034 OR "FID" =12142 OR "FID" =12199 OR "FID" =12601 OR "FID" =12655 OR "FID" =12813 OR "FID" =12902 OR "FID" =13081 OR "FID" =13081 OR "FID" =13271 OR "FID" =13271 OR "FID" =13362 OR "FID" =13457 OR "FID" =13527 OR "FID" =13564 OR "FID" =13620 OR "FID" =13817 OR "FID" =13916 OR "FID" =13922 OR "FID" =14115 OR "FID" =14159 OR "FID" =14291
N= 1904
Area = 50 km x 50 km

8. Picture content analysis
Big data
and maths
Picture content analysis
Method:
Observe all online pictures and classify them according to landscape elements and human activities
What cultural ecosystem services?
CHANGE: ADD MICRO PICTURES

9. Results Visual Experiential Intellectual
Landscape elements:
27% rural areas (mosaics of cropland, grassland, etc.)
18% water (river, lake, waterfall, etc.)
18% mountains and canyons
Etc.
Experiential
Activities:
14% recreation (hiking, relaxing, watching, biking, etc.).
Associated with elements (trails, rivers, lakes, etc.)
Intellectual
Heritage, legacy.
Spiritual, religious.
Inspiration and creativity.
Identity and sense of place.
Social relationships.
Education and knowledge. ??
This approach captures few services.
Focus on landscape beauty.
Some info on recreation.
Landscape pictures
27.0% rural areas (mosaics of cropland, grassland mosaic, and forest plantations)
18.2% water (river, lake, waterfall, etc.)
17.8% mountains and canyons
13.2% natural area (glacier, forest, grassland, etc.)
11.6% urban and buildings (main square, monument, etc.)
5.0% roads, bridges, tunnels
3.3% plants (flowers, trees, etc.)
+ Activities
14.3% recreation (hiking, relaxing, watching, biking, etc.)

10. Identifying who post pictures on Panoramio
Big data
and maths
Identifying who post pictures on Panoramio
Method:
Automatically download all pictures by all photographers having pictures in the region
Determine whether they take pictures locally (region), nationally (Peru) or internationally
Compare map density between local, national and international photographers

11. Results International Local
Limited information on actors. More national/international than local.
327 photographers with 95,211 posted pictures.
9% local, 32% national, 25% international.
Local
International
Different locations.
But strong effects of population and access.

12. Explaining picture density
Big data
and maths
Explaining picture density
Method:
Model the effect of landscape (land use, topography, hydrography, etc.) and human factors (accessibility, population density) on picture density
with Random Forest models ?

13. Results Good model performance
(explains 68% of variance)
Three best predictors of picture density:
For international photographers:
Distance to airport (accessibility)
Distance to protected area, Distance to wetlands (landscape)
For local photographers:
Population density, Distance to regional capital city (accessibility)
Distance to glacier (landscape)

14. Analyzing supply and demand of cultural services
Big data
and maths
Analyzing supply and demand of cultural services
Method:
Density shows service “use”
function of supply (by ecosystem)
and demand (by people)
Supply = model output without human factors
Demand = model output without natural factors ?
Supply
Demand
Use

15. Good approach for decomposing supply and demand
Results
Demand (by people)
Use (service)
Supply (by ecosystem)
Interna tional photo graphers
Supply by ecosystem depends on users: landscape beauty attributes are different for every person
Good approach for decomposing supply and demand
Local photo graphers

16. Fieldwork and social science
Asking local experts
Method:
Asking workshop participants to draw maps of where cultural services are used

17. Do experts know everything? Do they represent all views?
Results
Maps of location of the use of different services (hiking, “nice” forest landscapes, sacred places, etc.)
Do experts know everything? Do they represent all views?
No or few repetitions

18. Interviewing local people
Fieldwork and social science
Interviewing local people
Method:
Open interviews with people in the area (n=28)
With pictures (selected by interviewees) to prompt discussion
Discourse content analysis

19. This approach captures many services. Most cited are intellectual.
Results
People cited 40 different cultural ecosystem services and explained why
This approach captures many services. Most cited are intellectual.

20. Fieldwork and social science
Survey people
Fieldwork and social science
Method:
211 people, balanced between:
Urban / Rural / Tourists ; Men / Women ; Young (<20), Adults (20-40), Senior
6 pictures presented
For each, 8 statements on ecosystem services:
Recreation, Beauty, Purity, Identity, Peace, Tradition, Harmony and Production (not cultural)
Response on Likert scale
“full disagree” to “fully agree”
Linear Model:
Response as a function of picture, ecosystem service, and people

21. Results Many significant predictors of responses at p<0.01 >
Rural degraded
>
>
>
>
Rural
Lake
Grassland
Glacier
Beauty
> Peace, Recreation, Purity
> Harmony
> Tradition
> Identity
> Production
Urban
Also possible to build bundles
Non local > Local
Urban = Rural
Men = Women
Newcomer > Long-term resident
Old > Young people

22. Conclusion
Different approaches for assessing cultural ecosystem services for different research questions
Big data
Local fieldwork
Number of people studied: + ─ (cost)
Details on people: ─ +
Number of services studied: ─ +
Details on services: ─ +
Location of use, supply, demand: + ─
Drivers of use, supply, demand: + (quantitative) + (qualitative)

23. Thank you!