HAL Id: hal-03346749https://hal.archives-ouvertes.fr/hal-03346749

Submitted on 16 Sep 2021

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

A pollen-based land-cover atlas covering the last 200years of Bassies valley vegetation changes : results from

a new spatialization methodClara Plancher, Cédric Gaucherel, Didier Galop, Thomas Houet, L. Marquer,

S. Sugita, Florence Mazier

To cite this version:Clara Plancher, Cédric Gaucherel, Didier Galop, Thomas Houet, L. Marquer, et al.. A pollen-basedland-cover atlas covering the last 200 years of Bassies valley vegetation changes : results from anew spatialization method. International Symposium of Labex DRIIHM - 2021, Sep 2021, Toulouse,France. 2021, �10.34972/driihm-bcbd28�. �hal-03346749�

Figure 3 : Examples of pollen-based past land-cover maps

A) 1810 B) 1890 C) 1955 D) 1985

GrasslandsCalluna heathlands

Mixed heathlands

Rhododendron heathlands

Broadleaved forests

Mixed forests

Conifer forests

Peatbogs (constant)

A pollen-based land-cover atlas covering the last 200 years of Bassies valley vegeta�on changes : results from a new spa�alisa�on method Plancher C., Gaucherel C., Galop D., Houet T., Marquer L., Sugita S., Mazier F.

MATERIAL AND METHOD

CONTEXT

Interna�onal Symposium of Labex DRIIHM - 2021ANR-11-LABX-0010

6-8 September 2021- Toulouse (France)

0.50 1 KM

±

2 Spa�alisa�on models predic�on of past land-cover maps. Past + = Retrospec�ve land-cover models predic�ons and type selec�on at each pixel

Softwares : Multiscale Heterogeneity Map, ArcMap, RStudio

1 Spa�alisa�on models learning using sub-actual and (Fig.2).- Probability densi�es, spa�o-temporal autocorrela�on : Taxa distribu�on maps- Mul�ple linear modelling : = a + b + c , K-fold cross-valida�on

METHOD STEPS

OBJECTIVES

Esri, HERE, Garmin, (c)OpenStreetMap contributors,and the GIS user community

1800

800

1600

1200

1000

1400

2000

2200

2000

1800

2000

570000 575000

6185

000

0 10.5 KM

LEG

SIG

ESC

EM

W1652

F O U Z OT

A U Z A TV I C D E S S O S±

- Atlas of past land-cover over 200 years- New perspec�ves to study long-term rela�onship between land-use / land-cover and biodiversity

RESULTS

Landscape Reconstruc�on Algorithm es�matesPollen % Vegeta�on % of 20 dominant taxa in a 1 km radius around each site (Fig.1) for 10 to 20-year �me windows 5 topographic and hydrographic variables

Land-cover maps

1810

000.5

00.5

1

1.522.533.50

0.51

1.5

22.5

3

000.

51

1.5

22.5 0.5 1 1.5 0 0.50 0.5

11.50

0.50

1890

00

0.5

00.5

11.5

22.50.511.50

0.5

00.5

11.5

00

0.5

11.5

20 0.5 1 1.5 20 0.50 0

0.51

1.52

1945

1890 1945 1965

1965

2008

1810

1830

1850

1870

1890

1910

1930

1945195519651975198519952005

(Fig. 3C) (Fig. 3D)

00

0.5

00.5

11.5

200.511.52

00

0.51

1.52

000

0.5

1

1.52 2.5 3 0 0.5 1 0 0.50

0.51

1.52

0

000.

50

0.5

11.5

200.511.522.5

300

0.51

00

0.5

11.5

20 0.5 1 0 0.5 1 1.5 20

0.51

00.5

0

(km2)

(Fig. 3A,B) + Low rates of change (Fig.4)

Land-cover configura�on/composi�on Current biodiversity Past land-uses ? ?

- Grasslands declined by more than 50% over 200 years, mainly replaced by heathlands.- Phases of rates of change : (1) 19th century slowdown, (2) stabilisa�on un�l the 1940s, accelera�on in (3)

the 1950s and (4) from the 1970s.

CONCLUSION AND PERSPECTIVES

Figure 1 : Bassiès pollen sites : 2 small lakes ( ) and 5 bogs ( ) . Red line: Target area for land-cover reconstruc�on

1810

1870

1890

1910

1930

1945

1955

1965

1975

1985

1995

2005

1830

1850

19932008

TWMidpoint

Figure 2 : Dataset and �me windows. Environmental maps are considered constant over �me.

Environmental mapsFrom 1993 and 2008, 8 land-cover types

?

0.0 0.5 1.0 1.5 2.0

1810−1830

1830−1850

1850−1870

1870−1890

1890−1910

1910−1930

1930−1945

1945−1955

1955−1965

1965−1975

1975−1985

1985−1995

Interval

1995−2008

ChangesFastSlowUniform rate

Rate of change (%)Figure 4 : Rates of land-cover changes between �me windows midpoint

Have shown : But yet...Studies based on :

MacDonald et al., 2000. Journal of Environmental Management 59, 47–69 ; Houet et al., 2012. Sud-Ouest Européen 33, 41–56 ; Galop et al., 2011. PAGES News 19, 53–55 ; Herrault et al., 2015. Landscape Ecology 31, 877–893 ; Plancher et al., in prep.

No long-term AND spa�ally explicit reconstruc�on of past land-cover and

their legacy effect on biodiversity

Pastoral decline in european mountains

v Shrubs encroachment

Biodiversity decrease

Spa�ally explicit short term data (Aerial images, land-cover proxy)

ORLong-term data

(Palaeoecological data, vegeta�on and diversity proxy)

1,2

3,4

1 2 34 5

52

5

20m resolu�on

SITES AND DATASET