基于SWAT模型的流溪河流域土地利用与气候变化对径流的影响

Impact of changes in land use and climate on the runoff in Liuxihe Watershed based on SWAT model

选用国内外广泛应用的SWAT分布式水文模型,定量分析流溪河流域土地利用与气候变化对径流的影响,采用情景模拟分析方法设置3类情景进行定量分析.对上中下游的温泉、太平场和南岗3个水文站依次校正与验证得出:除温泉站在验证期的3个系数刚达标之外,其他的相对误差<15%、相关系数>0.8、Nash-Sutcliffe效率系数>0.75,说明SWAT模型在流溪河流域的径流量模拟中具有较高的适用性.综合型情景模拟分析得出:以1991—2000年为基准期,2001—2010年土地利用与气候变化综合引起年均径流量增加11.23 m3·s-1,土地利用变化引起年均径流量减少0.62 m3·s-1,气候变化引起年均径流量增加11.85 m3·s-1,气候变化的影响强度强于土地利用变化的影响强度.极端土地利用情景模拟分析得出:与2000年土地利用现状模拟径流量相比,耕地情景和草地情景的径流量分别增加2.7%和0.5%,林地情景的径流量减少0.7%,证明林地有一定的截流能力.气候变化情景模拟分析得出:流域径流量变化与降水变化呈正相关关系(降水每升高10%,径流平均增加11.6%),与气温变化呈负相关关系(气温每升高1℃,径流平均降低0.8%),降水变化的影响强度强于气温变化的影响强度.在气候变化环境下,需要重视对强降雨的预测和灾害预防,可通过优化土地利用结构与空间布局来减缓气候变化带来的水文负效应,如洪涝灾害.

SWAT model, an extensively used distributed hydrological model, was used to quantita- tively analyze the influences of changes in land use and climate on the runoff at watershed scale. Liuxihe Watershed＇ s SWAT model was established and three scenarios were set. The calibration and validation at three hydrological stations of Wenquan, Taipingehang and Nangang showed that the three factors of Wenquan station just only reached the standard in validated period, and the other two stations had relative error （RE）〈15%, correlation coefficient （R2） 〉0.8 and Nash-Suteliffe ef- ficiency valve （Ens）〉0.75, suggesting that SWAT model was appropriate for simulating runoff re- sponse to land use change and climate variability in Liuxihe watershed. According to the integrated scenario simulation, the annual runoff increased by 11.23 m3 · s-1 from 2001 to 2010 compared with the baseline period from 1991 to 2000, among which, the land use change caused an annual runoff reduction of 0.62 m3 · s-1, whereas climate variability caused an annual runoff increase of 11.85 m3 · s-1 Apparently, the impact of climate variability was stronger than that of land use change. On the other hand, the scenario simulation of extreme land use showed that compared with the land use in 2000, the annual runoff of the farmland scenario and the grassland scenario in- creased by 2.7% and 0.5% respectively, while that of the forest land scenario were reduced by 0.7%, which suggested that forest land had an ability of diversion closure. Furthermore, the scenar- io simulation of climatic variability indicated that the change of river runoff correlated positively with precipitation change （increase of 11.6% in annual runoff with increase of 10% in annual precipitati- on）, but negatively with air temperature change （reduction of 0.8% in annual runoff with increase of 1℃ in annual mean air temperature）, which showed that the impact of precipitation variability was stronger than that of air temperature change. Therefore, in face of climate variability, we need to pay attention to strong rainfall forecasts, optimization of land use structure and spatial distribu- tion, which could reduce the negative hydrological effects （such as floods） induced by climate change.