Modelling species niches and distributions: vision for the future Antoine Guisan

OB

Departement of Ecology and Evolution (DEE, FBM) Institute of Earth Surface Dynamics (IDyST, FGSE) ECEM 2014 – Ecological Modelling Beyond boundaries: next generation modelling Marrakech, 27-30 October 2014

1. Setting the scene

2

Titre de la présentation lundi 26 janvier 2015

Growing interest in predicting species distributions Species Distribution Models (SDMs)

0/

0000

of publications

Spectacular increase of SDM papers

year 3

Côté & Reynolds (2002) Science Graphe tiré de Guisan et al. (2013) Ecology Letters

Water

Species only occur in suitable abiotic conditions within their environmental niche, and these can be discontinuous in geographic space, not all accessible and not all biotically suitable

rainfall

Species have niches: biological data can hardly be spatially interpolated!

nichetemperature habitat duality!

fundamental niche

G. E. Hutchinson (1903-1991)

Temperature 4

N-dim

Hutchinson (1957) Cold Spring Harbor Symposia on Quantitative Biology Colwell & Rangel (2009) PNAS, Guisan et al. (2014) TREE

Principle of species distribution models (SDMs) Data collection

Statistical modelling

Spatial predictions

water

Field observations

…

…

Temperature Environmental maps

…

Model of the observed environmental niche

Potential distribution of the species

presence absence

5

Guisan & Zimmermann (2000) Ecological Modelling Guisan & Thuiller (2005) Ecology Letters

A major need for ecologically-meaningful environmental maps + slope + solar traject.

Exposure

Radiations

Altitude + meteo measures

Climate: T, P, ..

Slope

+ edaphic measures

Soils

6

e.g. www.unil.ch/rechalpvd

etc..

Virtually applicable to all organisms Guisan & Hofer (2003) J. Biogeography

Pellet et al. (2004) Conservation Biology

Frog

Patthey et al. (In review) J. Wildlife Management

Lütolf et al. (2006) J. Appl. Ecol.

Soil bacteria? Maggini et al. (2002) Biodiversity & Conservation

Jaberg & Guisan (2001) J. Appl. Ecol.

SDMs Soil fungi?

Maggini et al. (2006) J. Biogeography

Moretti et al. (2006) Petitpierre et al. Le Lay et al. J. Biogeography (2012) Science (2010) Ecogr. See Guisan & Thuiller (2005) Ecol. Lett., Elith & Leathwick (2009) AREES, Franklin (2010) CUP book

SDMs in fundamental sciences Biogeographic patterns

Spatial genetics

Species distributions

Environmental niche

Community assembly

Niche-habitat duality

Drivers of biological invasions

Dispersal/migrations pathways 8

Deriving projections in space or time 1A. modifying the input climate maps ∆ Temperature ∆ Landuse

? 2. reapplying the model

1B. providing input maps for a distinct study area

(i.e. the quantified niche)

New potential distribution in time or space 9

Guisan & Zimmermann (2000) Ecological Modelling Guisan & Thuiller (2005) Ecology Letters

Addressing global change issues Biological invasions

climate change

(projecting in space)

(projecting in time)

D. octopetala

E. myosuroides

present L. alpinus

Androsace

2030

Graphs reflect the rates of plant species extinctions in mountain ranges per elevation belt

 niche assumed to be ‘projectable’ in a new range or a new time period 10

Thuiller et al.(2005) GCB, Engler et al. (2011) GCB

SDMs in Conservation Sciences Discovering populations

Biodiversity monitoring

Anticipating invasions

Climate change impacts

Habitat restorations

Species reintroduction

Prioritization

Reserve selection

2. Vision for my future research at UNIL (DEE and IDYST)

12 Titre de la présentation lundi 26 janvier 2015

2.1 Methodological challenges - Novel developments needed - E.g. working more with “Virtual Ecologist” (VE) framework

GLM

- Improving tools (e.g. R libraries, ‘migclim’, ‘ecospat’)

13

Thibaud et al. (2014) Methods in Ecology & Evolution

2.2 Theoretical challenges Which niche are we modelling? SDMs are modeling the realized environmental niche not colonized

excluded by biotic interactions

14

available climate

Guisan et al. (20014) Trends in Ecology & Evolution

Effect of niche changes?

Realized niche in:

Range/time 1 Range/time 2

change in niche limit change in niche centroid reduced density of occurrences (but not empty)

15

Broennimann et al. (2012) Global Ecol. & Biogeogr. Guisan et al. (2014) Trends in Ecology & Evolution

In space

(7 spp, Past -6K to Present)

(50 invasive plants EU – US)

Picea

0.7

Fagus Abies

Dominant

0.6

Larix

0.5

Corylus

0.4

Pioneer regression line: r-square = 0.81 P = 0.00345

0

0.5

1.0

Unfilling Expansion

0.8

In time

Carpinus

0.3

Quality of past prediction

Niche changes lower SDM performances

niche overlap Juniperus

1.5

2.0

2.5

3.0

Intensity of niche change

16

Red: North-American spp invading Europe Green: European spp invading North-Am.

Pearman et al. (2008) Ecol. Letters Petitpierre et al. (2012) Science

Building the niche through time or space a1

a2

a3

Partial realized niches

a1+ a2+a3 t1+ t2 + t3

Pooled niche

a = area t = time CF = compositional factor (niche axis) fundamental niche realized niche

(close to fundamental?)

17

Maiorano et al. (2013) GEB, Modified from Nogues-Bravo (2009)

Beech, Fagus sylvatica

Model based on partial niches

(with current niche)

niche overlap

Model based on pooled niche

K years BP 18

Maiorano et al. (2013) GEB,

2.3 From species to communities

2 Center for Macroecology,

Evolution and Climate, University of Copenhagen Denmark

SDMs

Stacked SDM

of publications

2.4 SDMs to support decisions?

0/

000

Searching for applications in four fields of conservation

without ‘decision’ with ‘decision’

year

 Not so many used routinely! 20

Guisan et al. (2013) Ecology Letters

Putting SDMs in a decision context SDMs

Identifying the problem

Defining the objectives

SDMs

Defining possible actions

e.g. reserve selection options?

SDMs

Consequences of actions

e.g. impacts of translocation?

Uncertainty assessment

e.g. which exotic species may be problematic?

Trade-offs between costs and benefits of actions

Decision 21

Guisan et al. (2013) Ecology Letters

SDMs e.g. mapping confidence limits for predicted distributions?

The ever insufficient science-policy dialog Science

Conservation

• Scientists • Applied researchers • Modellers

Uncovering requirements and defining needs for synthesis

Scientific Knowledge and Tools

SDM Development and Evaluation

• Decision makers • Site managers • Practitioners

Synthesizing

‘Translators’ • Individuals • Groups • Institutions

22

Conservation Problems

Synthesizing

Model-assisted Decision-making

Guisan et al. (2013) Ecology Letters

2.5 Promoting multi-disciplinary research

23

Http://rechalpvd.unil.ch

Thanks for your attention

24 Titre de la présentation lundi 26 janvier 2015