[Effective Methods for Predicting Carbon Exchange in Desert Photovoltaic Ecosystems: Support Vector Machine Model...

Photovoltaic development （PVPPC） has gradually become an important way to address climate change and achieve energy transition. Under the influence of PVPPC， a unique photovoltaic ecosystem is formed by the interaction between biological communities and inorganic environments within the photovoltaic field. Maintaining carbon balance is crucial for achieving the sustainability and health of the photovoltaic ecosystem. Net ecosystem carbon exchange （NEE） helps measure the carbon cycle balance of photovoltaic ecosystems， which is influenced by various environmental factors such as meteorology and soil. Taking the Gonghe photovoltaic park on the Qinghai Tibet Plateau as the research area， field-measured meteorological， soil， and flux data were obtained to analyze the mutual feedback response relationship between ecological environmental factors and the NEE of desert photovoltaic ecosystems. It was found that net radiation， air temperature， wind speed， relative humidity， and average atmospheric pressure were the five driving factors that had the greatest impact on the NEE of the desert photovoltaic ecosystems. A support vector machine （SSA-SVM） optimized based on the sparrow search algorithm was used to construct an ecosystem NEE estimation model under the influence of desert photovoltaic development. The model was used to predict the changes in NEE of desert photovoltaic ecosystems under different climate scenarios. The results showed that the model had good simulation performance for the NEE of the desert photovoltaic ecosystem， with an error controlled within 2%. Under three climate scenarios （SSP126， SSP245， and SSP585）， the carbon sink of the desert photovoltaic ecosystem during the growing season was higher than that during the non-growing season. The average annual NEE （calculated as C） was -37.96， -41.32， and -47.68 g·（m²·a）^-1 and -12.69， -12.25， and -12.33 g·（m²·a）^-1. The impact of climate change on carbon cycling during the growing season was significantly higher than that during the non-growing season， indicating that the desert photovoltaic ecosystem will still maintain strong carbon sequestration potential in the future. This study provides a new perspective for predicting carbon exchange in desert photovoltaic ecosystems and also provides data support for fields such as ecosystem stability assessment， environmental restoration， and climate change trend analysis.