【目的】研究黄土高原植被覆盖对水热条件的时空响应。【方法】利用经验正交函数分析植被覆盖、气温和降水变化情况的时空分布,并结合奇异值分解分析植被覆盖与水热条件的时空相关性。【结果】东南部水热条件对植被的生长最佳,西北部干旱-半干旱地区气温超过一定的界限不利于植被生长,而降水的增加更容易使植被覆盖增大;中国第二大河——黄河的水文效应对干旱-半干旱地区植被覆盖的变化影响密切,在干旱地区水分充足的区域,气温的上升则会促进植被的生长;河谷平原、盆地中植被覆盖对水热的响应较为明显;同一区域不同植被类型对水热的响应也不尽相同;黄土高原植被覆盖与水热要素的相关性很强,但限制性因素较多。【结论】土地利用、植被类型、作物熟制以及地形地貌和高程等是黄土高原气温对植被覆盖影响的主要限制性因素。黄土高原地处干旱的内陆地区,降水对植被的影响最为直接。荒漠地带及非农业耕种区的植被覆盖状况主要取决于天然降水,两者之间具有很强的正相关性。农耕区由于受灌溉条件等人为因素的影响,植被覆盖和降水之间的相关情况更加复杂,不确定性更强。
【Objective】The vegetation response to temperature and precipitation in Loess Plateau was studied.【Method】EOF(Empirical Orthogonal Functions) and SVD(Singular Value Decomposition) were applied to analyze the temporal and spatial response characteristics of vegetation NDVI to the variation of temperature and precipitation,and to distinguish the spatio-temporal correlation between vegetation index and climate factors.【Result】The conditions for vegetation cover in southeast region of Loess Plateau were the best.More precipitation will be favourable to the vegetation cover,including the hydrological effect of Yellow River,but the temperature higher than one threshold value will be less good for plants in arid and semi-arid regions.Higher temperature made plants grow well where there is sufficient moisture in arid region.The vegetation cover has close relations to hydro-thermal condition in valley plain and basin valley.When there is this additional condition,the vegetation cover increases in hydro-thermal equilibrium region.There is a high level of correlation between vegetation cover and hydro-thermal condition,but more limiting factors.【Conclusion】The main limiting factors of temperature on the vegetation cover include land use,vegetation type,cropping system and landform.Loess Plateau is located in drought inland region,so precipitation has direct influence on vegetation cover.Plant growth depends on precipitation in desert zone and non-farming areas.The correlation between vegetation and precipitation is more complicated under irrigation conditions and so on in crop field.