Mapping land cover in complex Mediterranean landscapes using Landsat: Improved classification accuracies from integrating multi-seasonal and synthetic imagery

摘要: Abstract Low-intensity farming systems are of great importance for biodiversity in Europe, but they often affected by soil degradation or economic pressure, leading to either land abandonment intensification agriculture. These changes use influence local patterns and require annual monitoring cover. To accurately map cover such spatio-temporal complex landscapes, it is important capture their phenological dynamics fine spatial heterogeneity. Multi-seasonal analyses using optical sensors with a medium resolution from 10 60 m (e.g. Landsat) have been used this task, data availability can be scarce due cloud cover, sub-optimal acquisition schedules archive access restrictions. Combining coarse the MODerate-resolution Imaging Spectroradiometer (MODIS) Landsat provides opportunities close these gaps simulating Landsat-like images at MODIS temporal resolution. In study, we test whether what degree maps Mediterranean landscapes improve integrating multi-seasonal imagery, as well STARFM-simulated imagery whenever original observations unavailable. Therefore, develop different classification scenarios based on seasonally varying simulated data. Results show that spring early autumn crucial achieving satisfying mapping accuracies (overall accuracy 74.5%). Using synthetic increases compared single-date data, were never good We conclude essential STARFM compensate missing observations. However, if sufficient all phenologically dates, suggest relying solely Landsat.