Landscapes evolution have significantly altered the Earth's energy balance and biogeochemical cycles, thereby exacerbating climate change. This, in turn, affects surface characteristics and the provision of ecosystem services, especially carbon storage. While recent centuries have witnessed unprecedented landscape changes, limited long-term studies have offered insights into the comparison between present-day features and historical conditions. This study utilized historical reconstruction data and remote sensing imagery to assess landscape evolution and its consequences for carbon stocks over 300 years. Employing multiple regression and random forest models were selected to quantify the influence of key landscape metrics on carbon stocks in the Dongting Lake basin, allowing for a thorough analysis across different sub-basins and land types. The results revealed that intensified human disturbances led to increased landscape fragmentation (+82%), regularity (+56%), and diversity (+37%) within the basin. Moreover, carbon stocks decreased from 4.13 Gt to 3.66 Gt, representing an 11.4% loss, with soil carbon stock experiencing the most considerable reduction (0.24 Gt, 51%). These changes in carbon stock metrics corresponded to shifts in landscape patterns, both undergoing significant transitions at the turn of the 21st century. Meanwhile, fragmentation and regularity played a vital role in explaining carbon stock changes, as their increase contributes to greater carbon losses. Likewise, an increase in landscape diversity correlated with decreased carbon stocks, challenging the prevailing notion that enhanced diversity promotes carbon stocks. The influence of landscape patterns on carbon stocks varies notably across distinct land types. An increase in the dominance of farmland and built-up land led to decreased carbon stocks, while the opposite holds true for forestland. Similarly, a decrease in regularity for farmland, forestland, and built-up land benefits carbon storage, while grassland demonstrates the opposite trend. These findings offer insights for countries and regions in the early stages of development or approaching development, suggesting improvements in land use practices and strategies to address climate change. This involves offsetting land-based carbon emissions through changes in landscape spatial configuration.
Read full abstract