如何将基于站点观测的降水数据转化为分布式水文模型所需的空间分布降水数据, 是分布式水文模型模拟流 域水文过程的前提。椭圆指数函数模型具有能够指明降水中心位置及中心降水量的优点, 但与其他空间插值模型 比较, 椭圆指数函数模型的插值结果如何尚需进行比较研究。将椭圆指数函数模型与距离倒数权重模型、空间线性 插值模型、修正距离倒数模型等降水空间插值模型分别应用于Santa Cata lina 岛3 场次降水分布。插值结果表明, 椭圆指数函数模型和修正距离倒数模型由于考虑了更多影响因子和降水分布特性, 能较准确反映降水分布。空间 线性插值模型由于插值时只考虑相邻三个测点而忽略其他测点对插值结果的影响作用, 因而相比其他三种模型有 较大插值误差。修正距离倒数模型考虑了地形高程起伏影响因素, 其模拟精度优于距离倒数平方法。椭圆指数函 数模型由于具有连续分布函数特性, 能够模拟降水空间连续分布趋势, 因此与其他模型比较也有较高精度。

H ow to tr ansform the obser ved po int precipitatio n data at t he pr ecipitation stations into the distr ibuted data required by the hy dr olog ical model is the foundat ion f or the hy dr olog ical process simulation of distributed hydro lo gical model. T he Ellip2 tic Ex ponential Function M odel ( EEFM) has the capability o f identify ing the precipitation center po sitio n and center precipitat i2 o n amo unt. Compar ed w ith other spatial inter po lat ion models, EEFM inter po lation results remain unknow n and need further study. In this paper, Inver se Dist ance Weig hting ( IDW) , Space Linear I nter po lation M odel ( SLIM ) , Revise Inverse Distance Model ( RIDM) , and EEFM w ere a pplied to characterize t he distr ibut ion o f three precipitat ion events in t he Santa Catalina Is2 land. T he inter polat ion results show ed that ( 1) EEFM and RIDM are able to r eflect the pr ecipitation spatial distributio n mor e accurately due to that bo th metho ds consider mor e impact factors and precipitation distributio n char act eristics; ( 2) SLIM con2 siders the data from the t hr ee adjacent points only and ig nor es the effects o f o ther po ints, so its int erpolat ion results show lar ger erro rs compared to o ther models; ( 3) RIDM t akes into account the impact factor of the ups and downs of terr ain elevat ion, and its inter po lation accuracy is better than that of IDW; and ( 4) EEFM is a continuo us dist ribution functio n and it can simulate the co nt inuously spat ial distr ibution trend o f rainfall, therefor e it has a higher accuracy compar ed to o ther three mo dels.

国家自然科学基金项目( 40971021; 41471025) ; 山东省自然科学基金项目( ZR2014DM004)