Towards a new reconstruction of past regional land-cover in Central Asia
Résumé
Land cover and land use influence the climate system from regional to global scales through biogeochemical (e.g. CO2) and biophysical (e.g. energy exchanges in the biosphere) feedbacks. Quantitative information about these changes are critical for accurate vegetation and climate modelling in order to quantify their impacts on the climate system.
A long-term perspective (centuries to millennia) of changes in land cover and land use can be assessed by using pollen-based modelling approaches such as the REVEALS model. REVEALS estimates are now available for much of the northern Hemisphere. However, few data exist for the vast area in the centre of the Eurasian continent, the Central Asian steppes and piedmonts. Since this region is situated at the intersection between the major northern hemispheric climate subsystems, it is necessary to apply REVEALS to Central Asia in order to understand vegetation-climate feedbacks and dynamics on the world’s largest continent.
Central Asia is characterized by a variety of environmental (deserts/steppes to forests) and topographical (ranging between 3000-7500 m for the highest peaks) factors that affect pollen-vegetation relationships and thereby REVEALS estimates. Available pollen records in this region substantially overestimate pollen types deriving from lowlands and steppe-like environments, in particular Artemisia and Chenopodiaceae taxa. Relative pollen productivity estimates (RPPs) of these taxa have been difficult to assess so far and correspond to a vast range of values. The greatest challenge for RPPs in this region include the selection of sites and approach to vegetation collection; the inclusion of different species into a unique pollen type and landscape characteristics may play an additional role.
The present study explores the pollen-vegetation relationships of steppe- and desert-like plant taxa, focusing on Artemisia and Chenopodiaceae pollen types. The study area is located in the Ili River basin lowlands of southeast Kazakhstan. We undertook botanical survey, created vegetation maps and took modern pollen samples at thirty-seven randomly selected locations dominated by steppe- and desert-like vegetation within a radius of 55km. Here, we 1) discuss pollen taphonomy in terms of preservation, production, dispersion and deposition, 2) estimate the relevant source area for pollen and 3) calculate RPPs for the major plant taxa. We used the Extended R-value model coupled with the Lagrangian Stochastic Model (LSM) and Gaussian Plume Model (GPM) for modelling. RPP outcomes based on both LSM and GPM models are required to analyze the effects of wind-driven uphill transport and the differential vertical dispersion of pollen, as well as the long-distance transport of pollen in the mountain ranges of Central Asia, and therefore to get accurate land-cover reconstructions for this region of the northern Hemisphere.