作者: Robert Gilmore Pontius , Wideke Boersma , Jean-Christophe Castella , Keith Clarke , Ton de Nijs
DOI: 10.1007/S00168-007-0138-2
关键词: Resolution (logic) 、 Land change 、 Range (statistics) 、 Data mining 、 Raw data 、 Raster graphics 、 Figure of merit 、 Statistics 、 Computer science 、 Input/output 、 Null (SQL)
摘要: This paper applies methods of multiple resolution map comparison to quantify characteristics for 13 applications 9 different popular peer-reviewed land change models. Each modeling application simulates categories in raster maps from an initial time a subsequent time. For each application, the statistical compare: (1) reference time, (2) and (3) prediction The three possible two-map comparisons characterize: dynamics landscape, behavior model, accuracy prediction. three-map specifies amount prediction’s that is attributable persistence versus change. Results show error larger than correctly predicted 12 at raw data. are summarized compared using two statistics: null figure merit. According merit, more accurate ones where observed net larger. facilitates communication among modelers, because it illustrates range results variety models scientifically rigorous, generally applicable, intellectually accessible techniques.