Evrendilek Fatih
Department of Environmental Engineering, Faculty of Engineering and Architecture, Abant Izzet Baysal University, 14280 Bolu, Turkey.
Sensors (Basel). 2007 Dec 12;7(12):3242-3257. doi: 10.3390/s7123242.
This study aims at quantifying spatio-temporal dynamics of monthly mean dailyincident photosynthetically active radiation (PAR) over a vast and complex terrain such asTurkey. The spatial interpolation method of universal kriging, and the combination ofmultiple linear regression (MLR) models and map algebra techniques were implemented togenerate surface maps of PAR with a grid resolution of 500 x 500 m as a function of fivegeographical and 14 climatic variables. Performance of the geostatistical and MLR modelswas compared using mean prediction error (MPE), root-mean-square prediction error(RMSPE), average standard prediction error (ASE), mean standardized prediction error(MSPE), root-mean-square standardized prediction error (RMSSPE), and adjustedcoefficient of determination (R²). The best-fit MLR- and universal kriging-generatedmodels of monthly mean daily PAR were validated against an independent 37-year observeddataset of 35 climate stations derived from 160 stations across Turkey by the Jackknifingmethod. The spatial variability patterns of monthly mean daily incident PAR were moreaccurately reflected in the surface maps created by the MLR-based models than in thosecreated by the universal kriging method, in particular, for spring (May) and autumn(November). The MLR-based spatial interpolation algorithms of PAR described in thisstudy indicated the significance of the multifactor approach to understanding and mappingspatio-temporal dynamics of PAR for a complex terrain over meso-scales.
本研究旨在量化土耳其这样广阔而复杂地形上每月平均日光合有效辐射（PAR）的时空动态。采用通用克里金空间插值方法，结合多元线性回归（MLR）模型和地图代数技术，以生成网格分辨率为500×500米的PAR表面地图，该地图是五个地理变量和14个气候变量的函数。使用平均预测误差（MPE）、均方根预测误差（RMSPE）、平均标准预测误差（ASE）、平均标准化预测误差（MSPE）、均方根标准化预测误差（RMSSPE）和调整决定系数（R²）比较了地质统计模型和MLR模型的性能。通过留一法，利用来自土耳其160个站点的35个气候站的37年独立观测数据集，对月平均日PAR的最佳拟合MLR和通用克里金生成模型进行了验证。与通用克里金方法生成的表面地图相比，基于MLR的模型生成的表面地图更准确地反映了月平均日入射PAR的空间变异模式，特别是在春季（5月）和秋季（11月）。本研究中描述的基于MLR的PAR空间插值算法表明，多因素方法对于理解和绘制中尺度复杂地形上PAR的时空动态具有重要意义。
