DIFFICULT MIGRATION OF TEMPERATE TREE SPECIES

Transcription

1 UNIVERSITÉ DU QUÉBEC À RIMOUSKI APRIL 30, 2015 DIFFICULT MIGRATION OF TEMPERATE TREE SPECIES IN BOREAL FOREST UNDER CLIMATE CHANGE? Steve Vissault, Matthew Talluto, Isabelle Boulangeat and Dominique Gravel

2 CONTEXT THE BOREAL-TEMPERATE ECOTONE The surface of the boreal-temperate forests ecotone is expected to shift over the next 100 years. Goldblum and Rig, 2010

3 CONTEXT THE BOREAL-TEMPERATE ECOTONE 1. The location of this ecotone is responsive to climate. 2 Goldblum and Rig, 2010

4 CONTEXT PREDICTED FUTURE SPECIES DISTRIBUTION 2. Several temperate forest species are predicted to shift northward under climate change Sugar maple Red oak Yellow birch American ash 3 McKenney et al, 2007; Woodall et al, 2008; Iverson and Prasad, Web illustrations

5 CONTEXT PREDICTED FUTURE SPECIES DISTRIBUTION 2. Several temperate forest species are predicted to shift northward under climate change Future climate enveloppe of Sugar maple ( ) 4 McKenney et al, 2007

6 CONTEXT LIMITS AND DIFFICULTIES IN THIS STUDY CONTEXT Forest have: 1. Limited dispersions 2. Slow population dynamics 3. Interspecific competition 5 Travis et al, 2013

7 CONTEXT LIMITS AND DIFFICULTIES IN THIS STUDY CONTEXT Forest have: 1. Limited dispersions 2. Slow population dynamics 3. Interspecific competition These components will be affected by future climate 5 Travis et al, 2013

8 CONTEXT MODELLING COMPROMISE Spa al dynamic Explicit States and Transi on Models (STMs) Implicit Excluded Species Distribu on Models (SDMs) Excluded Implicit Explicit Demography/Interac on 6 Scheller and Mladenoff et al, 2006; Kriticos et al, 2013

9 STUDY OBJECTIVE Main objective: Assess range shift and migration rates of the temperate forest community toward boreal forest under climate change. 7

10 STUDY OBJECTIVE Main objective: Assess range shift and migration rates of the temperate forest community toward boreal forest under climate change. Why? Predict the future distribution of temperate species community in Quebec Improve and adapt our forests management practices under climate change 7

11 NEW APPROACH STATES AND TRANSITIONS MODEL B R T Model Description Lanscape scale 4 States: T, Temperate B, Boreal M, Mixed R, corresponds to a post-disturbance Spatially explicit and stochastic model M 8

12 NEW APPROACH STATES AND TRANSITIONS MODEL R Disturbance Ecological processes: Disturbance B Disturbance Disturbance T M 9

13 NEW APPROACH STATES AND TRANSITIONS MODEL Succession R Succession Ecological processes: Disturbance Succession B Succession T M 10

14 NEW APPROACH STATES AND TRANSITIONS MODEL R Ecological processes: Disturbance Succession Colonization B Colonization Colonization T M 11

15 NEW APPROACH STATES AND TRANSITIONS MODEL B R Ecological processes: Disturbance Succession Colonization Competitive exclusion T Exclusion M Exclusion 12

16 NEW APPROACH STATES AND TRANSITIONS MODEL Ecological processes: R Disturbance Succession Colonization Competitive exclusion B T Each probability depends on: Proportion of states available in the neighborhood M Local climatic conditions (Precipitation, Temperature) 13

17 DATA THE QUICC-FOR DATABASE Forest permanent plots databases USA Ontario Québec New-Brunswick Temporary tables F(x) Final relational database Postgres 9.3 QUICC-FOR Climatic data ANUSPLIN 14

18 CALIBRATION DATA USED 1. Classify state of each plot Plot remeasured Transition observed between remeasurements 15

19 CALIBRATION DATA USED Temperate Classify state of each plot Boreal Plot remeasured Transition observed between remeasurements Latitude Number of plots (10,50] (50,100] (100,200] (200,300] (300,400] (500,800] Mixed Longitude

20 CALIBRATION DATA USED Temperate Classify state of each plot Boreal Plot remeasured Transition observed between remeasurements 2. Compute state transition probabilities based on the actual climate and neighbors plot states. Latitude Mixed Number of plots (10,50] (50,100] (100,200] (200,300] (300,400] (500,800] Longitude

21 CALIBRATION TRANSITION PROBABILITIES OVER CLIMATIC GRADIENTS R B T 10 M Temperature ( C) 5 0 Class of probability (0,0.06] (0.06,0.12] (0.12,0.18] (0.18,0.24] (0.24,0.3] (0.3,0.36] (0.36,0.42] (0.42,0.48] (0.48,0.54] (0.54,0.6] Precipitation (m) 16

22 SIMULATIONS PREDICT THE NEXT TIMESTEP Cell area:1 ha Ini al state: Temperate Neighborhood (8 cells) Annual mean temperature (1 cell) Annual precipita on (1 cell) State at the next mestep? 17

23 RESULTS PREDICT THE CURRENT LANDSCAPE Current states distributions predicted on climatology: Latitude State B M R T Longitude 18

24 RESULTS PREDICT THE CURRENT LANDSCAPE Temperature increase linearly (4 C): 2100 Latitude State B M R T Longitude 19

25 RESULTS MIGRATION RATE PREDICTED Distance (km) Proportion of temperate states (T+M) Years:

26 SUMMARY 1. Spatial dynamique, demography and species interactions constraint temperate migration 2. Slow temperate migration rate predicted by the STM 3. Tension between potential and realized distribution (at equilibrium with climate) 21

27 AKNOWLEDGEMENTS Funded by In collaboration with 22

28 CALIBRATION EXTRA SLIDES Step 2. Simulated anenaling {}}{ P(D t1 M t0, Climate) = f( Climate, ˆD, ˆM ) (1) }{{} Step 1. RandomForest 23

29 BIOCLIMATIC DOMAINS EXTRA SLIDES Landscape predicted by the STM Bioclimatic domains in Quebec Latitude State B M R T Longitude 24

30 PROXY OF THE NEIGHBOORS: SDM EXTRA SLIDES 25