青藏地区1 km牲畜数据集（1544-2020）

长时间序列牲畜数据集的研制是理解人类放牧活动的气候和生态效应基础，对于实现人与自然和谐共生具有重要意义。基于历史文献记载、代用指标和统计资料得到青藏地区1544-2020年省域牲畜数量（羊单位），采用地理探测器遴选部分自然、人文影响因子构建放牧适宜性模型，利用地理加权随机森林（GRF）机器学习方法模拟牲畜密度分布空间格局构建了一套1km×1km的1544-2020年牲畜数据集。本研究估算的青藏地区牲畜数量变化趋势与该地区畜牧业兴衰过程较为吻合，空间重建结果与草地利用强度空间分布与变化呈现较好的一致性，表明本研究重建结果较可靠。目前长时间序列放牧活动定量重建几近空白，本研究有望为青藏地区牧业管理和其他牧区历史放牧强度定量重建提供参考。

Development of long-time-series livestock datasets is fundamental to understanding the climatic and ecological impacts of human grazing activities, and holds great significance for achieving harmonious coexistence between humans and nature. Based on historical documents, proxy indicators and statistical data, the provincial-scale livestock quantity (sheep equivalent unit, SEU) of the Qinghai-Tibet Region from 1544 to 2020 was compiled. Subsequently, partial natural and anthropogenic influencing factors were screened using Geodetector to establish a grazing suitability model, and the spatial pattern of livestock density distribution was simulated via the machine learning method of Geographically Weighted Random Forest (GRF). Ultimately, a 1km × 1km livestock dataset spanning the period from 1544 to 2020 was constructed. The trend of livestock quantity in the Qinghai-Tibet Region estimated in this study aligns well with the rise and fall of local animal husbandry, and the spatial reconstruction results exhibit high consistency with the spatial distribution and changes of grassland use intensity, confirming the reliability of the reconstruction results in this study. At present, quantitative reconstructions of long-time-series grazing activities are nearly non-existent, and this study is expected to provide a valuable reference for livestock husbandry management in the Qinghai-Tibet Region and quantitative reconstruction of historical grazing intensity in other pastoral areas.

搜集汇总

背景与挑战

背景概述

青藏地区1 km牲畜数据集（1544-2020）是一个长时间序列的牲畜数据集，涵盖了1544年至2020年的青藏地区牲畜数量（羊单位），空间分辨率为1km×1km。该数据集基于历史文献记载、代用指标和统计资料，采用地理加权随机森林（GRF）机器学习方法模拟牲畜密度分布空间格局，为青藏地区牧业管理和其他牧区历史放牧强度定量重建提供参考。