ABSTRACT
Land cover is a fundamental earth-surface attribute shaped by geologic, hydrologic, climatic, atmospheric, and land-use processes occurring at a range of space-time scales. Land cover, in turn, affects these processes through feedback mechanisms such as plant respiration, which both absorbs and releases carbon, water, oxygen, and other biochemical elements from or to the environment. Therefore, knowledge of land cover is essential to understand earth-surface processes relevant for managing land and preserving natural environments. Examples include climate and weather modeling (Bonan, 2004), carbon budget assessment (Turner et al., 2004), water supply and quality analysis (Chang, 2003), evaluation of terrestrial and aquatic ecosystem integrity (Eshleman, 2004; Fraser et al., 2009), investigation of effects of farming practices on erosion and on nutrient contamination
12.1 Introduction .......................................................................................................................... 177 12.2 Implementing Classi‡cation for Land-Cover Time Series ................................................... 178
12.2.1 Classi‡er Retraining ................................................................................................. 179 12.2.2 Classi‡er Extension .................................................................................................. 179
12.3 Factors Affecting Land-Cover Time Series Accuracy and Temporal Consistency.............. 180 12.3.1 Classi‡cation Methods .............................................................................................. 180 12.3.2 Training Data ............................................................................................................ 181 12.3.3 Thematic Resolution and Separability ...................................................................... 181 12.3.4 Data Consistency ...................................................................................................... 183
12.4 Additional Processing to Improve Temporal Consistency ................................................... 183 12.4.1 Expert Rules ............................................................................................................. 183 12.4.2 Fuzzy Reasoning....................................................................................................... 184 12.4.3 Transition Matrices ................................................................................................... 185 12.4.4 Change Area Constraints .......................................................................................... 186
12.5 Examples of Land-Cover Time Series .................................................................................. 186 12.5.1 Global........................................................................................................................ 186 12.5.2 Australia .................................................................................................................... 186 12.5.3 Canada ...................................................................................................................... 187 12.5.4 United States ............................................................................................................. 187
References ...................................................................................................................................... 188
in lakes and rivers (Potter, 2004), evaluation of suitability of land for infrastructure development or biofuel production (Fischer et al., 2010), wildlife population modeling (Kerr and Ostrovsky, 2003), and biodiversity assessment (Kerr, 2001).
