Influences of changing land use and CO2 concentration on ecosystem and landscape level carbon and water balances in mountainous terrain of the Stubai Valley, Austria

Abstract A process-based spatial simulation model was used to estimate gross primary production, ecosystem respiration, net ecosystem CO 2 exchange and water use by the vegetation in Stubai Valley, Austria at landscape scale. The simulations were run for individual years from early spring to late fall, providing estimates in grasslands for carbon gain, biomass and leaf area development, allocation of photoproducts to the below ground ecosystem compartment, and water use. In the case of evergreen coniferous forests, gas exchange is estimated, but spatial simulation of growth over the single annual cycles is not included. Spatial parameterization of the model is derived for forest LAI based on remote sensing, for soil characteristics by generalization from spatial surveys and for climate drivers from observations at monitoring stations along the elevation gradient and from modelling of incident radiation in complex terrain. Validation of the model was carried out at point scale, and was based on comparison of model output at selected locations with observations along elevation gradients in Stubai Valley and Berchtesgaden National Park, Germany as well as with known trends in ecosystem response documented in the literature. The utility of the model for describing long-term changes in carbon and water balances at landscape scale is demonstrated in the context of land use change that occurred between 1861 and 2002 in Stubai Valley. During this period, coniferous forest increased in extent by ca. 11% of the vegetated area of 1861, primarily in the subalpine zone. Managed grassland decreased by 46%, while abandoned grassland and natural alpine mats increased by 14 and 11%, respectively. At point scale, the formulated model predicts higher canopy conductance in 1861 due to lower atmospheric CO 2 concentration which opens stomata. As a result, water use at point scale decreased by ca. 8% in 2002 in the valley bottoms versus 10% at tree line. At landscape level, the decrease in water use by vegetation in 2002 was predicted to be twice as high (ca. 17%) due to increase in subalpine forest, reduction of managed grassland in the valley and on slopes, as well as abandonment of grassland which results in natural succession. Net ecosystem CO 2 exchange (NEE) was predicted to increase (become more negative) at point scale depending on vegetation type by 10 to 20% in 2002 due to increasing atmospheric CO 2 concentration. However, due to the shift from grassland to forest and natural vegetation, landscape level CO 2 exchange did not change. As a result of land use change, the export of carbon in harvested biomass in 2002 was estimated at only 30% of that in 1861. While the need for further validation of model assumptions is recognized, especially changes in ecosystem behavior with changing atmospheric CO 2 concentration, the model analysis indicates a long-term reduction in water use by vegetation and a shift in ecosystem services. The results provide a case study, where land use change may compensate or override the influences of increasing atmospheric CO 2 concentration, maintaining a relatively constant NEE in present time period simulations as compared to 1861, as well as reducing export of carbon from the alpine landscape of Stubai Valley. Use of the model in evaluation of scenarios of future land use change and in relation to vulnerability of ecosystem services are discussed.