We want you to know what is going on in the BOD, our meetings, our actions, members leaving, the new ones elected,... but text written in this blog cannot be taken an official position or statement of the Society for Conservation Biology. Probably it is not even an official statement of the section... as these need to be approved by the members.

Wednesday, 28 November 2012

Long-term vegetation dynamics in wood-pastures simulated with WoodPaM

Wood-pastures are inherently dynamic systems consisting of a mosaic of patches of forest, grassland and interconnecting successional stages of shrub land (see e.g. Vojta in this blog). Grazing and browsing of livestock negatively impacts tree regeneration and counteracts the natural successional trend towards forest development (in a Middle European context). The balance of these processes, and therefore the landscape mosaic, however, is very sensitive to external influences, such as forest management and climate change. Both directly impact the recruitment and growth of tree species, as well as the spatial configuration of grassland and forest patches, which feedbacks to successional dynamics and grazing behavior of free ranging livestock. Such feedback mechanisms complicate predictions of the effect of management decisions. Moreover, the long-term dynamics at landscape level due to the slow establishment and growth of trees temporarily decouple management actions from the long-term system’s response.

In order to cope with this problem of spatio-temporal complexity in wood-pasture management, a simulation model of wood-pasture ecosystems WoodPaM was developed, which primarily aimed on a strategic investigation of the dynamics and feedback mechanisms in wood-pastures in the Swiss Jura Mountains (Figure 1, Gillet et al. 2002, Gillet 2008, see Buttler et al. in this blog). Recent model refinements now allow a more tactical application to investigate the potential fate of wood-pastures in the light of climate and land-use changes.

Figure 1. Structure of the WoodPaM-model and main interactions (arrows) among submodels (in horizontal direction) and across spatial scales (in vertical direction). Successions in the herb, the shrub and the tree submodels are driven by grazing impacts and climate. Grazing intensity at the grid cell level is determined by selective habitat use of cattle at the landscape level. Local successions in turn feed back to landscape level by the formation of spatio-temporal patterns of forage availability and landscape structures. Figure 1 is modified after a manuscript by Peringer et al. currently in review in Ecology and Society.
In a case study (details in Gillet & Peringer 2012) we simulated the vegetation dynamics (grassland succession, tree establishment, growth and decay) in the wood-pasture La Bullatonne Dessous, which is situated in the Jura Vaudois at high elevation from 1225 to 1480 m a.s.l. (Coordinates 46.856083 / 6.564403), for two climate change scenarios in combination with three scenarios of forest management. The climate change scenarios pinpoint two extreme possible futures of our world, a fuel-intensive future with drastic warming (IPCC-SRES 2000 A1FI, +6 K from 2000 to 2100) and a moderate development with less warming (IPCC-SRES 2000 B2, +4 K from 2000 to 2100). Forest management strategies consider: (i) no management (NM); (ii) assisted migration (AM), assuming an increased immigration of 20 seedlings of each tree species in each grid cell per year during the warming period 2001-2100; (iii) forest management (FM), assuming logging every 20 years of 25% of the trees in randomly selected cells that contain at least 20 trees of the same species. The six resulting simulation experiments explore the options to adapt management practices to different degrees of climate warming and are subsequently named B2-NM, B2-AM, B2-FM, A1FI-NM, A1FI-AM and A1FI-FM. Simulations are prolonged until calendar year 3000 to investigate long-term successional trends from hypothetical equilibrium patterns that are triggered by today’s management decisions under upcoming climatic conditions.

The simulations show that climate-change impacts are delayed for decades and centuries after warming started in calendar year 2001, and that future landscape dynamics and structures strongly depend on the degree of warming (Figures 2 and 3). After minor changes until 2050, divergent succession lines lead to progressive forest succession following moderate warming (B2-NM), while extreme warming temporarily triggers regressive succession (A1FI-NM, Figure 2 a,b). This is explained by the drought-induced collapse of currently dominant Norway spruce (Picea abies) and its slow replacement by beech (Fagus sylvatica) or by Scots pine (Pinus sylvestris) dependent on the degree of warming. In the long run, dominance of beech triggers a segregation of open grassland and closed forest, while dominance of pine leads to a homogenous landscape consisting of densely wooded pastures (Figure 3). Assisted migration of tree species in our AM scenarios does not change the final configuration of the landscape but reduces the time required for species replacement, thus preventing the breakdown of current spruce forests (Figure 2 c,d). Logging turned out to homogenize the landscape, if dominated by beech in the moderate climate change scenario (B2-FM, Figure 3) through the prevention of the development of any grazed forests or unwooded pastures (Figure 2 e). While in the moderate climate change scenario (B2-FM) a stable state developed after centuries (2600), in the pine dominated landscape of the extreme warming scenario (A1FI-FM), logging triggered shifting-mosaic dynamics (Figure 2 f). Here, landscape complexity was subsequently enhanced  (Figure 3). 

Figure 2. Landscape dynamics in the wood-pasture „La Bullatonne“ following a combination of two climate change scenarios (moderate B2 and extreme A1FI in columns) and three forest management scenarios (no management a and b, assisted migration c and d, patchwise clearing e and f in rows). Landscape structure is displayed in terms of the cover of four phytocoenosis-types inside the pasture, which are mainly defined by tree cover: unwooded pastures with tree cover below 1%; sparsely wooded pastures with tree cover between 1% and 20%, trees and shrubs being mostly scattered; densely wooded pastures with tree cover within the range 20-70% and a coarse-grained structure; grazed forests with treecover higher than 70% and small clearings included in a forest matrix. This figure is from Gillet & Peringer (2012).

Figure 3. Time series of the landscape aggregation index (He et al. 2000) following two scenarios of climate change (B2 left and A1FI right) and three strategies of forest management. A higher landscape aggregation index means a more homogenous landscape. Figure corresponds to Figure 3 in Gillet & Peringer (2012).

We conclude that in contrast of historical studies, which show a somewhat resilience of the wood-pastures of the Jura to past climate variability (Sjögren 2006), the resilience and adaptive capacity will be challenged in the future due to inexorable changes in tree species composition and landscape structure driven by warming and drought stress. However, our simulations clearly outline the potential of forest management to alleviate these effects of climate change. Assisted migration smoothes tree species shift and logging counteracts the successional trend triggered by climate change. However, the same logging strategy may lead to either relative simplification (B2-FM) or complexification (A1FI-FM), dependent on the tree species composition. Thus, any direct human intervention should be carried out carefully and after considering the current state of the system together with observed and expected successional trends.

Gillet, F. & A. Peringer. 2012. Dynamic Modelling of Silvopastoral Landscape Structure: Scenarios for Future Climate and Land Use. 2012 International Congress on Environmental Modelling and Software, International Environmental Modelling and Software Society (iEMSs), Leipzig, Germany. http://www.iemss.org/society/index.php/iemss-2012-proceedings.
Gillet, F. 2008. Modelling vegetation dynamics in heterogeneous pasture-woodland landscapes. Ecological Modelling 217:1-18.
Gillet, F., O. Besson, and J.-M. Gobat. 2002. PATUMOD: a compartment model of vegetation dynamics in wooded pastures. Ecological Modelling 147:267-290.
He, H. S., B. E. DeZonia, and D. J. Mladenoff. 2000. An aggregation index (AI) to quantify spatial patterns of landscapes. Landscape Ecology 15:591-601.
IPCC. 2000. Emissions Scenarios - Summary for Policymakers. A Special Report of IPCC Working Group III. Intergovernmental Panel on Climate Change.
Sjogren, P. 2006. The development of pasture woodland in the southwest Swiss Jura Mountains over 2000 years, based on three adjacent peat profiles. Holocene 16:210-223.

Prof. François Gillet
Community ecology and dynamic modelling
Université de Franche-Comté – CNRS, France

Dr.-Ing. Alexander Peringer
Institut für Landschaftsplanung und Ökologie ILPOE
Universität Stuttgart Germany

1 comment:

bromsgrove bathrooms said...

Catalytic types use and a combustor and makes it easier to clean and maintain. A non catalytic re-circulates the ash and deposits a part of it on the stove making regular cleaning necessary.