Landscape Pattern, Fire Regime And Vegetation Dynamics

Transcript

Centro de Estudios Ambientales del Mediterráneo, València, Spain Landscape pattern, fire regime and vegetation dynamics - A modelling approach Juli G Pausas MEDECOS 2004, Rhodes The Mediterranean basin „ Landscapes of the MB can only be undertood considering: z z Millenary history of human population Millenary history of intense land use Mosaics: fields and wildland Oldfields: changes in landscape pattern 1935 Montseny, NE Spain (wet Medit.) 1981 1910 Vall d’Albaida, E Spain (dry Medit.) 2002 Recent changes Socioeconomic changes -> Landscape changes -> Fire changes 80 70’s 19th Valencia region, E Iberian Peninsula 1994-00 1984-93 1974-83 1968-73 1954-61 1944-53 1934-43 Year 1924-33 1980 1914-23 1960 1904-13 1940 1894-03 1920 1884-93 1900 1874-83 0 1 Agriculture Industry & services 0 Number of fires 10 10000 100 15000 Area burned (ha/yr) Number of fires (#/yr) 5000 Annual area burned (ha/yr) 60 40 20 Active population (%) 500 Fire history Active population 20th Pausas 2004, Clim. Change 63 Climate change Summer temperature 24 23 22 Mean summer temperature (°C) 100 80 60 40 20 Summer rainfall (mm) 120 25 Summer rainfall 1950 1960 1970 1980 1990 2000 Years Valencia region, E Iberian Peninsula 1950 1960 1970 1980 1990 2000 Years Pausas 2004, Clim. Change 63 Questions „ To what extent landscape pattern matters: 1 - Does different landscape configurations determine different vegetation dynamics? 2 - Does different landscape configurations determine different fire regime? Approach „ Changes at long-term: Modelling • hypothesis generator tool „ Functional Types • workable simplification • 2 fire-based FT x 2 life forms „ Simulations • 5 Landscape scenarios: L1 .. L5 • 6 Fire regime scenarios: f0 .. f5 „ Results z Fire regime x landscape pattern Fire-based Funtional Types Obligate Resprouters (R+P-): „ Tree: Quercus FT (Q. ilex) „ Shrub: Erica FT (E. multiflora) Obligate Seeders (R-P+): „ Tree: Pinus FT (P. halepensis) „ Shrub: Cistus FT (C. albidus) The model „ FATE (Moore & Noble 1990) z z z z z z „ based on few functional attributes developed for fire-prone ecosystems has been tested for Mediterranean conditions simple, yet capturing main details on veg dynamics individual cohort-based model however, it is not spatial FATELAND - landscape version of FATE z gridded FATE + dispersal + fire propagation 5 Landscape scenarios L1 „ „ „ „ L2 L3 L4 L5 100 x 100 cells of 10 x 10 m Distribution of 4 FTs (Q, E, P, C) Covering ca 30% each L1 - L5: 5 different degrees of patchiness (aggregation) Quercus Cistus Pinus Richness Dominant Coarser L1 Erica L2 L3 L4 Finer L5 Landscape pattern gradient 80000 Total edge Quercus Erica Cistus Pinus Quercus Erica Cistus Pinus 40000 20000 200 300 Total edge (m) 60000 Quercus Erica Cistus Pinus 0 0 100 Number of patches 400 500 Number of patches L1 L2 L3 L4 L5 L1 L3 L4 Richness Autocorrelation L5 Autocorrelation 0.8 0.6 0.4 Autocorrelation (I) 150 100 0.2 50 20 15 L1 L2 Coarse L3 L4 L5 Fine L1 L2 Coarse L3 L4 L5 Fine 0.0 0 5 10 Quercus Erica Cistus Pinus 0 Mean Patch Size (ha) 25 Quercus Erica Cistus Pinus Spatial Autocorrelation (JC z) 30 Mean patch size L2 L1 L2 L3 L4 L5 6 Fire scenarios f0 0 „ „ „ „ „ f80 f40 0.0125 0.025 f20 f10 f5 0.05 0.1 0.2 Stating in time: Annual fire prob.: 0 - 0.2 Starting in space: uniform randomly Propagation: Probabilistic fuel-dependent 500 yr simulation: fire regime analysis Year = 100: Landscape analysis Simulation results (preliminary) Initial F0 F80 F40 F20 F10 F5 L1 L2 L3 L4 L5 Cistus (yr= 100) Initial F0 F80 F40 F20 F10 F5 L1 L2 L3 L4 L5 Pinus (yr= 100) Landscape attributes: Edge f0 initial Resprouters Total edge (km) Total edge (km) 100 f80 Quercus f40 f20 f10 f5 100 Pinus 80 80 60 60 40 40 20 20 0 100 0 100 Erica Cistus 80 80 60 60 40 40 20 20 0 0 L1 Coarse L2 L3 L4 L5 L1 Fine L2 Coarse L3 L4 Seeders L5 Fine yr= 100 Abundance Coarser, aggregated Seeders Resprouters L1 f0 Finer L2 L3 L4 L5 Q E C P f80 f40 f20 f10 f5 500 years simulation Fire regime Coarser, aggregated Finer f10 f20 f40 f80 10000 9500 f5 f10 f20 f40 f80 f5 f10 f20 f40 f80 0.8 f5 8500 9000 9500 8500 9000 9500 9000 8500 Fire size L5 10000 L3 10000 L1 0.6 0.4 0.0 0.2 Spatial autocorrelation of fire recurrence f80 f40 f20 f10 f5 f5 f10 f20 f40 f80 L1 L2 L3 L4 L5 Conclusions „ Landscape pattern effect landscape processes 1. The dynamics and the abundance of a plant species depend on the initial spatial configuration. And the differences are stronger when fire frequency is high (interaction effect) 2. Fire regime depends of the initial spatial configuration „ FATELAND as a tool for fire-landscape interactions in Mediterranean conditions New hypotheses „ „ „ Aggregated patterns (H autocor) -> slower dynamics Fire-sensitive species may be maintained with aggregated patterns (i.e. fragmentation accelerates extintion of fire-sensitive species) Heterogeneity of fire history is related to the aggregation of vegetation Need to be verified Thanks! Ευχαριστίες