Grain for Green Project dominates greening in afforested areas rather than that in grass revegetation areas of the Loess Plateau, China—using Deep Crossing LSTM Age network

作者：Xueting Wang, Honglin He, Mengyu Zhang, Jianming Deng, Xiaoli Ren, Yan Lv, Weihua Liu, Zining Lin, Shiyu Dong · 发表于：Environmental Research Letters · 年份：2025 · DOI：10.1088/1748-9326/adec02 · 被引用次数：6 · 研究领域：Plant Ecology and Soil Science、Forest, Soil, and Plant Ecology in China、Rangeland and Wildlife Management

Abstract Vegetation restoration in the Loess Plateau (LP) of China is driven by atmospheric environmental changes (climate change, rising CO 2 , and nitrogen deposition), land cover change (LCC) from ecological restoration projects (ERPs), and change in forest age. However, the dominant factors influencing vegetation restoration remain controversial. This study improved the Deep Crossing network by integrating bidirectional long short-term memory (Bi-LSTM) with embedding, creating the Deep Crossing LSTM Age (DC-LSTM-Age) network. It incorporates land cover type, forest age, and atmospheric environmental factors to reconstruct the leaf area index (LAI). We investigated the LAI increase (greening) driven by various factors and their dynamics in the Grain for Green Project (GGP) regions of the LP from 2001 to 2021. Results showed that DC-LSTM-Age network effectively simulated LAI values and its temporal dynamics in LCC regions, with superior validation performance ( R 2 = 0.87) compared to the Deep Crossing LSTM network ( R 2 = 0.84) that excluded forest age and the Bi-LSTM network ( R 2 = 0.79) that excluded forest age and land cover type. The greening trend in afforested regions (GGP-Forest, 0.013 m 2 m −2 yr −1 ) was much larger than in grass revegetation regions (GGP-Grass, 0.005 m 2 m −2 yr −1 ). Dominant drivers varied by restoration strategy: in GGP-Forest, LCC was the primary driver (0.25 m 2 m −2 , 52.9%), with an increasing impact over time. In GGP-Grass, atmospheric e...