This study was designed to examine relationships between climate and vegetation of the Pacific rim of North America, from the Mediterranean deserts of California to Alaska’s boreal taiga. Relations were inferred from...This study was designed to examine relationships between climate and vegetation of the Pacific rim of North America, from the Mediterranean deserts of California to Alaska’s boreal taiga. Relations were inferred from temperature and rainfall data recorded at 457 weather stations and by sampling the vegetation around these stations. Climate data were used to construct climatograms, calculate forty one variables and detect main latitudinal and longitudinal gradients. In order to identify the best functions able to relate our variables, polynomial and non-polynomial regressions were performed. The k-means algorithm was the clustering method used to validate the variables that could best support our bioclimatic classification. The variable that best fitted our classification was finally used to prepare a discriminatory key for bioclimates. Across this extensive area three macrobioclimates were identified, Mediterranean, Temperate and Bo- real, within which we were able to distinguish nine bioclimates. Finally, we relate the different types of potential natural vegetation to each of these bioclimates and describe their floristic composition and physiognomy.展开更多
基金supported by grants from the Franklin Institute of North American Studies(University of Al-calá).
文摘This study was designed to examine relationships between climate and vegetation of the Pacific rim of North America, from the Mediterranean deserts of California to Alaska’s boreal taiga. Relations were inferred from temperature and rainfall data recorded at 457 weather stations and by sampling the vegetation around these stations. Climate data were used to construct climatograms, calculate forty one variables and detect main latitudinal and longitudinal gradients. In order to identify the best functions able to relate our variables, polynomial and non-polynomial regressions were performed. The k-means algorithm was the clustering method used to validate the variables that could best support our bioclimatic classification. The variable that best fitted our classification was finally used to prepare a discriminatory key for bioclimates. Across this extensive area three macrobioclimates were identified, Mediterranean, Temperate and Bo- real, within which we were able to distinguish nine bioclimates. Finally, we relate the different types of potential natural vegetation to each of these bioclimates and describe their floristic composition and physiognomy.
