[Objective] The variation characteristics of non-uniformity for the distribution of annual runoff of Chishui River and its response to precipitation was analyzed in order to provide a basis for decision-making regarding the development and utilization of water resources and flood control in the upper and middle reaches of the Chishui River basin. [Methods] We constructed a database for the SWAT model to simulate watershed runoff. The model was calibrated and validated with measured monthly runoff data. Model output data were combined with precipitation/runoff concentration degree (PCD/RCD) and concentration period (PCP/RCP), and the characteristics of annual runoff distribution and its response to precipitation were determined. [Results] The coefficient of determination (R2) and Nash-Sutcliffe efficiency coefficient (Ens) values for two hydrological stations were greater than 0.83 for the calibration period, and greater than 0.69 for the validation period, which meet the accuracy requirements. The annual distributions of precipitation and runoff was significantly uneven, and the change trends for the two were relatively consistent, with precipitation and runoff mainly concentrated in June to August. The temporal and spatial distributions of RCD were significantly affected by PCD. Due to infiltration and evapotranspiration, watershed RCD was usually greater than PCD. But when PCD<0.3, RCD was no longer dominated by PCD. Due to the lag in the response of watershed runoff to precipitation changes, RCP was often greater than PCP, and a small, short-term increase in precipitation had a significant impact on PCP, but a limited effect on RCP. [Conclusion] Precipitation was the leading factor causing changes in RCD and RCP in Chishui River watershed. Changes in the runoff coefficient in different periods were the main reason for the different response characteristics of RCD/RCP compared with PCD/PCP.

喀斯特地区地形起伏大,常规的降尺度方法及所选择的因子对其不适用.该文根据喀斯特地区的特点,选取反射率、遥感指数及高程因子为尺度因子,通过随机森林模型建立MODIS第31、32波段辐射亮度数据和尺度因子之间的非线性关系,构建适合喀斯特地区的随机森林(Karst Random Forest,KRF)模型,成功将空间分辨率为1 km的热红外辐射亮度降至100 m,最后利用劈窗算法反演得到空间分辨率为100 m的地表温度.将KRF方法与仅考虑常规因子的多因子随机森林回归模型(MTVRF)和热锐化算法(TsHARP)对比,结果表明:1)在不同高差的喀斯特地区,KRF方法可较大程度提高地表温度降尺度精度,均方根误差(RMSE)在遵义市西北部和贵阳市以南地区分别为2.46 K和1.45 K,较MTVRF模型分别降低了0.1419 K和0.2928 K,较TsHARP算法分别降低了0.6204 K和0.6953 K,且在地形起伏度较低的喀斯特山区城市(贵阳市以南)表现更好;2)在喀斯特地区不同地类上,KRF方法效果也较好,其中植被区域最优,RMSE为1.41 K,破碎的裸土区域RMSE为1.84 K.研究显示,考虑特殊尺度因子的KRF方法可提高喀斯特地区地表温度的降尺度精度,为该地区以地表温度为基础的研究提供更精细可靠的地表温度产品.

The effect of acid and aluminum stress on seed germination and seedling growth of Pinus massoniana Lamb.was studied on the substrates of quartz sand by compounding nutritive liquid commonly used intensively in the world.The result showed that medium intensity of acid and Al3+ did not have significant influence on germination rate and germinating ability,while they had significant negative influence on the growth of seedling root and stem.There was a significant correlation between the growth of root and stem,dry weight,fresh weight with intensity of acid and Al3+ stress.The seedling growth increased slowly at first and then decreased with pH value changing from 5.5 to 3.0,the seedling growth decreased with the increase of Al3+.The seedling growth was optimal in the pH value of 4.5～5.0.The acid and Al3+ stress had the biggest influence on root,and root was the most sensitive part.The percentage of biomass of root decreased with the increasing stress of acid and Al3+,both stem and leave fluctuated lesser.When pH value was under 4.5 of nutritive liquid,the concentration of Al3+ was 50 mg·L-1,which could restrain the growth of P.massoniana sprout distinctly.