预测不同放牧策略对草原土壤的影响

Predict the Effects of Different Grazing Strat-egies on Grassland Soils

摘要: 本文对不同放牧策略对植被生物量、土壤湿度、土壤有机物含量进行评价分析，从降水量、土壤蒸发等数据进行时间序列预测土壤湿度，从草原监测点历史监测土壤数据建立拟合预测不同放牧强度下的土壤有机物含量预测，得出一个放牧的最优策略，促进了草原的可持续发展。首先，从机理分析的角度建立熵权Topsis综合评价模型，分别评价不同的放牧策略对植被生物量和不同放牧强度对土壤湿度的影响。然后，使用基于时间序列的随机森林预测模型，预测2022年4月到2023年12月不同深度的土壤湿度。最后，从机理分析的角度出发建立基于专家权重的Topsis综合评价法，评价不同放牧策略对土壤化学性质的影响，然后结合附件相关数据，使用基于时间序列的多层LSTM模型进行预测在不同放牧强度下的2022年同期的一些有机碳、无机碳、全N和C/N比等值。

Abstract: In this paper, different grazing strategies evaluate and analyze vegetation biomass, soil moisture and soil organic matter content, predict soil moisture in time series from precipitation and soil evaporation data, establish a fitting prediction of soil organic matter content under different graz-ing intensities from the historical monitoring soil data of grassland monitoring points, and obtain an optimal grazing strategy, which promotes the sustainable development of grassland. Firstly, from the perspective of mechanism analysis, a comprehensive evaluation model of entropy weight Topsis is established, and the effects of different grazing strategies on vegetation biomass and different grazing intensities on soil moisture are evaluated respectively. Then, using a time-series-based random forest forecasting model, soil moisture at different depths from April 2022 to December 2023 is forecasted. Finally, from the perspective of mechanism analysis, the Topsis comprehensive evaluation method based on expert weighting is established to evaluate the effects of different grazing strategies on soil chemical properties, and then combined with the relevant data of attach-ments, a multi-layer LSTM model based on time series is used to predict some organic carbon, inor-ganic carbon, total N and C/N ratios in the same period of 2022 under different grazing intensities.