Analisi della vegetazione naturale in aree vulnerabili alla desertificazione mediante telerilevamento: i casi di studio di Monte Coppolo e Bosco Pantano in Basilicata
Author(s)
Trotta, Claudia
Date Issued
October 3, 2008
Type
Doctoral Thesis
Abstract
Southern Italy regions have been experiencing soil and vegetation degradation processes under the stress induced by both climate change and increased human activities whose effects are related to desertification processes. Land degradation generally result in a substantial change in vegetation patterns and distribution that can be profitably documented through remotely sensed data. Indeed, satellite sensor
technology provides a powerful tool that allows assessment and monitoring of vegetation covers and
biophysical parameters, with various levels of accuracy and with various scales (space and temporal), as well as providing a support to the development and execution of natural resources management plans.
In particular, this thesis focused on the study of vegetation by means of remote sensing data supported by reference data sampling, bibliographical and cartographic data to describe and spatialize vegetation distribution and ecological data. This study has been carried out in two sites, Monte Coppolo and Bosco Pantano, both localized in areas of Basilicata region susceptible to desertification, the activity has been articulated in three parts as described below.
In the first part, very high spatial resolution data (Ikonos) has been used in order to describe with a great detail the present distribution of vegetation canopies by means of a semi- automatic classification methodology which resulted competitive with traditional vegetation mapping techniques. Collected field data and satellite-derived vegetation maps, useful in the following phases of this study, overcome a gap
in the knowledge of the two areas and can constitute a reference point for further analyses. To further understand the vegetation changes occurred during the last five decades in relation to the modifications induced by human activity, historical aerial photos (1954) have been compared to recent ones (1998) and vegetation distribution modification has been evaluated.
In the second part, it has been used an integrated approach between high and medium spatial resolution data (Ikonos and Landsat) in order to implement semi-empirical models for retrieving an important biophysical parameter of the vegetation such as the leaf area index (LAI) by means of spectral reflectance data.
The aim was to test a method that exploits high resolution data (4 m) in describing the great
heterogeneity of LAI distribution as well as optimising field measurements for the calibration of the models. LAI maps derived by means of these models have been used like “ground truth” for the
calibration of models based on medium resolution data (30 m). Medium resolution data turns out useful in studying and monitoring those phenomena that assume importance on large-scale because, on the contrary of high resolution data, they are characterized by a greater extension of the sensed surface and by a regular acquisition, representing a valid tool for mo nitoring and studying territorial dynamics.
Medium resolution remote sensing data, i.e. Landsat, provide also a relatively wide historical series (approximately thirties years) which can be used to carry out multitemporal studies. Therefore, the performance and the stability (and hence the integrated approach) of the implemented semi-empirical models has been evaluated in order to test the possibility to temporarily extrapolate the found relations.
Finally, a time series of Landsat imagery (1984-2005) has been used in order to derive a multi-temporal series of spectral vegetation indices to be correlated to the inter-annual climatic variability as well as analyzing the different answers of the single vegetation formations in the study areas to variations of temperature and rainfall across years.
The achieved results show that remote sensing is a powerful tool to obtain qualitative description of vegetation distribution with great details and high accuracy in patchy and heterogeneous areas such as Mediterranean ones; conversely quantitative retrievals of biophysical parameters by means of regressive models are influenced by several factors such as phenological time of data acquisition.
Additional information
Dottorato di ricerca in Ecologia e gestione delle risorse biologiche
File(s)![Thumbnail Image]()
Name
ctrotta_tesid.pdf
Size
3.33 MB
Format
Adobe PDF
Checksum (MD5)
e6cb070713e9b83fc680b026a0d1e273