Response of European forests to extreme climatic events predicted for the 21 st century: sensitivity to climate models and their variability

Transcription

1 Response of European forests to extreme climatic events predicted for the 21 st century: sensitivity to climate models and their variability M. DURY, A. HAMBUCKERS, P.WARNANT, A. HENROT, E. FAVRE, M. OUBERDOUS and L. FRANCOIS Département d Astrophysique, de Géophysique et d Océanographie, Université de Liège Bât B5c,17 allée du Six Août, B- Liège (Belgium) Marie.Dury@ulg.ac.be COST Action FP93, Rome, 5-7 October, 21

2 Objectives Analyse the interannual variability of 4 climate models - comparison with present-day climate - impacts on ecosystems What will be the impact of climate change (mean and variability) on hydrological budget,, forest net primary productivity and forest fires? use of the CARAIB dynamic vegetation model forced with 4 climatic scenarios for the future

3 The CARAIB Model

4 CARAIB Dynamic Vegetation Model

5 Simulation strategy Fully transient simulations over Europe with 1-day 1 time step with CRU TS 3. historical climate data with - /Climate climate projections - -RACMO2 - -HIRHAM5 - Hadley Centre-HadRM3Q - interpolated to regular grid of.5 x.5 x.5 - only anomalies are used with respect to period - month to month variability from climate models - day to day variability from stochastic generator

6 CARAIB and present-day climate () )

7 CRU temperature and precipitation for Temperature ( C) Winter Summer > < -15 Precipitation (mm) Winter Summer >

8 CARAIB on present-day climate Model runoff Present Runoff from UNESCO (estimated from river discharge) mm/yr > Area burned in Spain and in Portugal over the period from CARAIB and from forest fire statistics Correlation coefficient =,72 JRC, 27. GPP (g C m 2 day -1 ) Gross primary productivity for Vielsam CarboEurope site ( N, 6 6 E) over the period Year Observed GPP CARAIB GPP

9 Present-day variability Models vs observations

10 Monthly precipitation (mm) 42 N 13,5 E (Italy) CRU Precipitation (mm) Year Precipitation (mm) Year Monthly soil water content (SW-WP/FC WP/FC-WP) WP) Soil water content (relative unit) 2 1,8 1,6 1,4 1,2 1,8,6,4, Soil water content (relative unit) 2 1,8 1,6 1,4 1,2 1,8,6,4, Year Year Annual net primary productivity (g C m - 2 yr - 1 ) NPP (g C m -2 yr -1 ) Year CRU

11 Winter temperature standard deviation ( C) CRU C >

12 Summer precipitation standard deviation (mm) CRU >

13 Annual NPP standard deviation (g C m - ² yr - 1 ) CRU g C m-² m >

14 Annual area burned (ha) Present Present CRU ha Present > Present Present % burned with respect to CRU Southern Europe Central Europe Northern Europe -CLIMAT -RACMO2 -HIRHAM5 -HadRM3Q HadRM3Q

15 Projections over A1B SRES

16 Winter and Summer temperature anomalies between and WINTER SUMMER c > <

17 Winter and Summer precipitation anomalies between and WINTER SUMMER mm > < -12

18 Annual soil water content anomalies between and % > < -8

19 Annual soil water variability for and relative units >

20 Annual NPP variability for and g C m-² m >

21 Annual area burned for Annual area burned in the Mediterranean Region over the period from CARAIB Area burned (ha) Year ARP CLIMAT -RACMO2 -HIRHAM5 Hadley Centre-HadRM3Q ha > Hadley Centre Increasing factor between and Southern Europe 2 2,1 1,4 2,4 Central Europe 3,1 8,7 3,8 4,1 Northern Europe 4 1,3 2,6 6,3 Mean 2 5 3,5

22 Conclusion The 4 climate models do not reproduce exactly the present natural climate variability This variability has a strong impact on modelled soil water, forest NPP and fires For the future, climate models predict important climate change in mean and in variability: - Soil water stress will be more frequent and more severe in southern and central Europe - NPP interannual variability will increase - Frequency and intensity of (natural) fires will increase in many parts of Europe

23 Annual net primary productivity for progressive CO2 increase (A1B SRES scenario) constant CO2 (no CO 2 fertilisation) % > < -8

24 Winter temperature variability for and >

25 Summer precipitation variability for and >