陕北黄土高原气象要素对植被覆盖的空间分异影响及风险探测

Spatial differentiation effects and risk detection of meteorological elements to vegetation cover on the Loess Plateau of northern Shaanxi

陕北黄土高原地处我国西北生态环境脆弱区,近年来受退耕还林(草)影响,区域植被覆盖变化显著,与气象要素的响应关系也呈现出了复杂的空间分异性和不确定性,亟待厘清。以不同地貌分区为响应单元,探究了陕北黄土高原变绿前后(1982—1992年与2005—2015年)NDVI指数的时空演化特征,分析了植被覆盖变化与气象要素之间的分异性响应关系,并透过最优参数的地理探测器模型(OPGD)探测了气象要素变化对植被覆盖演变的交互影响及风险。研究结果表明:(1)陕北黄土高原及各地貌分区在研究期内NDVI指数均呈显著上升趋势,且阶段化分异明显;各地貌区气温和实际蒸散发量在高原变绿前后趋势变化差异明显,1982—1992年主要呈不显著增加趋势,而2005—2015年则表现为不显著减少趋势;两个时期降水量的空间异质性较为显著,其中北部地区为持续增加趋势,而南部地区则由减少逆转为增加趋势。(2)不同地貌区NDVI指数与气温和蒸散发均呈正相关,且蒸散发与NDVI的相关性强于气温;相对高原变绿前,2005—2015年NDVI指数对各气象要素的响应强度均有所增大。(3)OPGD因子检测结果显示5km空间网格是评价气候变化对陕北黄土高原植被覆盖变化影响的最佳空间尺度,各地貌分区对NDVI指数变化影响最大的气象因子依次为:气温(黄土塬)、降水(盖沙黄土丘陵)、蒸散发(黄土峁状丘陵)、降水(黄土梁状丘陵)、降水(黄土宽谷丘陵)、蒸散发(风沙丘陵)、降水(土石丘陵);交互探测表明气象因子间的交互作用对NDVI指数的空间分异具有协同增强性,但不同时期各地貌区的主导交互因子略有差异。(4)风险探测发现不同实际蒸散发量对NDVI指数的影响有显著空间差异。(5)陕北黄土高原的植被覆盖变化并不能简单归因于退耕还林(草)等生态修复措施,而应是多要素耦合驱动的结果,气象要素在其中扮演了不可忽视的作用。

The Loess Plateau in northern Shaanxi is located in the ecologically vulnerable region of northwest China.In recent years,affected by the return of farmland to forestry(grassland),the regional vegetation cover has changed significantly,and the response relationship with meteorological elements has also exhibited complex spatial differentiation and uncertainty,which urgently needs to be clarified.Taking different geomorphic regions as response units,this study investigated the spatiotemporal evolution characteristics of the Normalized Difference Vegetation Index(NDVI)before and after the greening(1982—1992 and 2005—2015)of the Loess Plateau in northern Shaanxi,analyzed the differential response relationship between vegetation cover changes and meteorological elements,and explored the interaction effects and risks of meteorological elements changes on vegetation cover evolution through the optimal parameters-based geographical detector(OPGD)model.The results showed that:(1)the NDVI of the Loess Plateau in northern Shaanxi and each geomorphic subdivision presented a significant upward trend during the statistical period,with significantly phased differentiation.The trends in temperature and evapotranspiration in different geomorphic regions showed significant differences before and after the greening of the plateau,with a mainly non-significant increasing trend from 1982 to 1992 and a non-significant decreasing trend from 2005 to 2015.The spatial heterogeneity of precipitation was more significant during the two periods,with the northern region showing a continuously increasing trend,while the southern region reversed from a decreasing trend to an increasing trend.(2)The NDVI of different geomorphic regions was positively correlated with temperature and evapotranspiration,and the correlation between evapotranspiration and NDVI was stronger than that between temperature and NDVI.Compared with the pre-greening period,the response intensity of the NDVI to various meteorological elements increased from 2005 to 2015.(3)The OPGD factor detection results showed that a 5km spatial grid was the optimally spatial scale for evaluating the impact of climate change on vegetation cover change on the Loess Plateau in northern Shaanxi.The meteorological factors that had the greatest impact on the NDVI in different geomorphic regions were in order:temperature(Loess tableland),precipitation(Cover sand loess hills),evapotranspiration(Loess hilly mounds),precipitation(Loess beam hills),precipitation(Loess wide valley hills),evapotranspiration(Wind-sand hills),and precipitation(Soil-stone hills).The interactive detection showed that the interaction between the meteorological factors had a synergistic enhancement effect on the spatial differentiation of the NDVI,but the dominant interaction factors in different geomorphic regions differed slightly at different times.(4)The risk detection found significantly spatial differences in the influence of different evapotranspiration amounts on the NDVI.(5)The change in vegetation cover on the Loess Plateau in northern Shaanxi could not simply be attributed to ecological restoration measures such as the return of farmland to forestry(grass)but should be the result of a coupling of multiple factors,in which meteorological elements played a non-negligible role.