Search

Time series analysis plays an important role in the detection of mechanisms that drive population fluctuations. However, long time series are rare, with ungulate data sets usually not exceeding 50 years. In this article, we describe a long-term data set of population density indices of five ungulate species obtained from the analysis of bag records collected in the Castelporziano Preserve, Rome, Italy. Hunting statistics are often used as proxies for population density; however, in the case of long time series for large mammals, there are no comparative studies to assess the validity of such data. We evaluated the ungulate time series, using two different approaches: we 1) compared hunting statistics with independent animal counts, and 2) assessed whether or not habitat composition of the drive areas was representative of habitat availability in the whole estate. Regressions between bag data and animal counts gave significant results only for three species, whereas bag data corrected for hunted a...

In this paper a study carried out by airborne hyperspectral data is presented. Image data concern a forest area (60 km North of Rome -Italy), to evaluate how different spatial resolutions can affect vegetation spectral response and therefore the discrimination among the different communities. MIVIS images were acquired at different flight altitudes (2000 m and 5000 m), in the same day and on the same surface targets. A radiometric field survey was carried out in order to radiometrically calibrate the airborne images. Classification were performed on each image by two different techniques: Maximum Likelihood and Spectral Angle Mapper. The results of these classification methods were analysed to evaluate how different spatial resolution can affect vegetation spectral response. In particular the relationship between spectral and spatial resolution of hyperspectral images was investigated by resampling the 2000 m image (4m/pxl) in order to simulate the radiometric response of surface targets in the MIVIS image acquired at 5000 m of altitude (10m/pxl). The result suggest that the spatial resolution of aerial images must be decided when the overflight is planned, because the resampled data keeps the original spectral characteristics and upscaling methods do not provide meaningful data in heterogeneous area.