基于MCE-CA-Markov和InVEST模型的伊犁谷地碳储量时空演变及预测

陆地生态系统碳源与碳汇的变化与土地利用/覆被变化（LUCC）的演变密切相关，为探讨土地利用变化对陆地生态系统碳储量的影响机制，基于1980—2020年LUCC数据集，通过多评价准则（Multi-criteria evaluation，MCE）的元胞自动机（Cell automata，CA）和马尔科夫链（Markov chain）模型，模拟伊犁谷地2030年的LUCC时空动态，耦合InVEST模型探讨土地利用变化下伊犁谷地1980—2030年陆地生态系统碳储量的时空演变格局。结果表明，MCE-CA-Markov模型预测LUCC数据集与2000、2010年和2020年实际LUCC精度检验的Kappa系数分别为0.929 1、0.875 5和0.929 7，模型模拟普适性较高，可对后续碳储量时空演变格局进行精准评估。利用InVEST模型估算的伊犁谷地1980年总碳储量约为1 114.95 Tg，预计至2030年总碳储量呈逐期下降趋势，累计净减少65.94 Tg，其中林草地面积的退缩是致使碳储量下降的主导因素。伊犁谷地碳储量空间分布总体上表现为高值区域环绕低值区域，呈嵌套分布。碳密度的高值区域分布在南部和北部山区林草地，而低值区域仅集中在中部河谷平原附近。研究土地利用变化对陆地生态系统碳储量的影响可为研究区碳库管理提供数据支持，为制定碳固存和环境保护政策提供理论参考。

Temporal and spatial evolution and prediction of carbon stocks in Yili Valley based on MCE-CA-Markov and InVEST models

Changes in carbon sources and sinks in terrestrial ecosystems are closely related to the evolution of land use/cover change (LUCC). The mechanisms of land use change on carbon stocks in terrestrial ecosystems were investigated to provide theoretical references to understand the influence of spatio-temporal evolution patterns on regional carbon stocks. Based on the LUCC data of 1980-2020, the spatio-temporal dynamics of LUCC of the Yili Valley in 2030 were simulated through multi-criteria evaluation(MCE), cellular automata (CA), and Markov chain models. InVEST model explored the spatio-temporal evolution pattern of carbon stocks in the Yili Valley terrestrial ecosystem from 1980 to 2030 under the land use change. The results showed that the MCE-CA-Markov model predicted LUCC data set and the actual LUCC in 2000, 2010, and 2020, and their Kappa coefficients were 0.929 1, 0.875 5, and 0.929 7, respectively, indicating that model simulation was highly generalized and could be used to accurately assess the spatio-temporal evolution of carbon stocks in subsequent years; the InVEST model estimated that the total carbon stock of the Yili Valley was 1 114.95 Tg in 1980, and it would show a decreasing trend until 2030, with a cumulative net decrease of 65.9 Tg. The retreat of forest and grassland area was the dominant factor leading to the decrease in carbon stock; and the spatial distribution of carbon stocks in the Yili Valley was generally characterized by a nested distribution of high value areas around low value areas. The high value areas of carbon density were distributed in the southern and northern mountainous forests and grasslands, while the low value areas were concentrated in the central valley plains. Studying the impacts of land use change on the carbon stock of terrestrial ecosystems can provide data support for carbon pool management in the study area, and subsequently, provide theoretical reference for the formulation of carbon sequestration and environmental protection policies.