Multi-scale sampling and statistical linear estimators to assess land use status and change

摘要: Question: Multi-temporal analysis of remotely sensed imagery has proven to be a powerful tool for assessment and monitoring landscape diversity. Here the feasibility assessing land-use diversity change was tested at multiple scales over time by means statistical linear estimators based on probabilistic sampling design. Location: The study area (the district Asciano, Tuscany, Italy) is characterized erosional forms typical Pliocene claystone (i.e. calanchi biancane) that have been subject phenomenon biancane reworking past 50 years, mainly owing expansion intensive agriculture. Methods: Cells two different (50 m� m 10 m) were classified operators according multilevel legend, using 1954 2000 aerial photographs. Inter-operator agreement accuracy Cohen’s K coefficient. Total land cover estimation each class carried out multistage estimator, while variance estimated Wolter estimator. Field-based information plant species composition recorded in order test relationship between use community ANOVA indicator analysis. Results: Agreement photointerpreters significantly higher than those expected chance, proving approach proposed reproducible, as long proper quality assurance methods are used. Our data show that, considered m), crops increased against woodlands semi-natural areas, latter showing highest mean richness. Meanwhile, an increase coverage trees shrubs found within probably result secondary succession occurring elements such biancane. Conclusions: Inferential statistics made it possible acquire quantitative abundance classes, allowing formal multi-temporal multi-scale Sampling design-based trends considering both expenditure other costs. They make maintain same scale series detect coarse- fine-grained changes spatial patterns.