Change detection in land cover GIS databases. LCCS and CLC.

Olteanu, Vlad Gabriel ; Badea, Alexandru ; Dana, Iulia Florentina 等

1. INTRODUCTION

Land cover is the observed (bio) physical cover of the Earth's surface. Land use is characterized by the arrangements, activities and inputs people undertake in a certain land cover type to produce, change and maintain it (Di Gregorio & Jansen, 2000).

The GlobCover Classification system was created by the European Space Agency (ESA) in partnership with UN FAO and it is based on ENVISAT's Medium Resolution Imaging Spectrometer (MERIS) with a resolution of 300 m. GlobCover's thematic legend is compatible with LCCS (Arino et al., 2007). This system does not satisfy our needs for classification, due to its low spatial resolution. It is, however, a big step in global classification system and may be used in the future.

At national level, there are already several land cover geodatabases created that refer to Romania's territory and we intend to use them in detecting large scale changes that took place over the years.

The Land Cover Classification System (LCCS) was initiated by the United Nations Environment Program (UNEP) and the Food and Agriculture Organization (FAO) in 1993. The project intends to develop a set of rules and methods for data processing and validation that can be applied to images acquired once every three or four years. In the case of Romania two geodatabases are available and a third one, LCCS 2007, is now being created. The satellite images were taken from Landsat TM and have a resolution of 30 m.

Corine Land Cover Classification System was financed by the European Union and it comprises 44 classes grouped into 3 hierarchic levels. The images came from the Landsat ETM and have the same resolution of 30 m. (http://dataservice.eea.europa.eu/dataservice/metadetails.asp7id =1007).

2. METHODOLOGY

2.1 LCCS 2000--LCCS 2003

In the frame of LCCS Romania the two geodatabases are divided into modules that contain classes. The eight modules are represented by Cultivated and Managed Terrestrial Areas, Cultivated Aquatic Areas, Natural Terrestrial Vegetation, Natural Aquatic Vegetation, Urban and Artificial Surfaces, Bare Areas, Artificial Water Bodies, and Natural Water Bodies. In order to detect changes in the LCCS database percentages of occupied area were calculated for each module in both 2000 and 2003 geodatabases.

The results reveal that the Cultivated and Managed Terrestrial Areas increased by 0,71 %, the percentage of the Cultivated Aquatic Areas remained the same (0,28 %), the area of the Natural Terrestrial Vegetation decreased by 0,84 %, while the Urban and Artificial Surfaces increased by 0,17 %. Furthermore, the surfaces represented by bare areas decreased with 0,15 %, the Artificial Water Bodies increased by 0,03 % and the Natural Water Bodies decreased by 0,05 %.

The class represented by forests (F), included in the Natural Terrestrial Vegetation module, was calculated separately. The results show that the area covered with forests had a decrease of about 0.2% from 2000 to 2003. This decrease (from 28,7% to 28,5%) is most probably the result of deforestation. In Fig.1 an example of deforestation is presented. The first image presents the situation in 2000 for an area covered by a large forest near the city of Zalau. The second image presents the situation in 2003 for the same area. There are visible differences regarding the large forest in the middle due to deforestation.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

The increase in the Urban and Artificial Surfaces (0,17 %) is due to the intense construction works as most major cities are expanding. After the year 2000, Romania began a period of high urban development and therefore significant changes take place every year; thus small size herbaceous fields (SSH) or grasslands (GRL) near the urban areas are turned into built-up areas. These changes were easily detected in the geodatabases.

The changes presented above imply human activity (deforestation, urban development) but there are natural changes that can be detected as well. An example is given in the images in Fig. 2 which presents the evolution of the Sacalin sandy island in the Danube Delta from 2000 to 2003. Initially, there were two islands, but in time they have joined into only one, due to alluvial deposits. Presently, the island is the subject of intense costal erosion.

The chart in figure Fig 3 presents the relative increase or decrease for the evaluated modules: Cultivated and Managed Terrestrial Areas, Natural Terrestrial Vegetation, Natural Aquatic Vegetation, Urban and Artificial Surfaces, Bare Areas, Artificial Water Bodies and Natural Water Bodies. The Cultivated Aquatic Areas module was not represented given the fact that it had no relative increase comparative to the situation in the year 2000.

2.2 CLC 2000--2006

In order to validate our results obtained from the LCCS 2000 and 2003 geodatabases, the same information was extracted from the Corine Land Cover (CLC) geodatabase. The two legends are different and therefore we have reorganized the classification in CLC to match the classification in LCCS. Due to the fact that the CLC legend makes no difference between the Artificial Water Bodies and the Natural Water Bodies those modules were grouped in LCCS, obtaining only one module, that of Water Bodies.

[FIGURE 3 OMITTED]

CLC consists of two geodatabases, Corine 2000 and 2006. The results obtained by comparing these two databases are similar to the ones previously obtained with LCCS. A special comparison was made regarding deforestation and it revealed a decrease in the areas covered with forests of 0,2%.

3. CONCLUSION

The Land Cover/Land Use knowledge plays nowadays an important role in most of the domains regarding the environment: hydrology, land management, climate change, natural disasters, monitoring and use of natural resources (Lambin & Geist, 2006), etc. and a change detection approach in this case was necessary especially to assess the human impact on the environment and to predict future developments in certain areas.

The RO-LCCS-2007 project, which follows the GMES actions in Romania (Badea et al., 2008), will provide a great opportunity to evaluate the latest land cover changes--an important task, keeping in mind the high rate of development in the 2003--2007 period.

4. REFERENCES

Arino, O.; Gross, D.; Ranera, F.; Bourg, L.; Leroy, M.; Bicheron, P.; Latham, J.; Di Gregorio, A.; Brockman, C., Witt, R.; Defourny, P.; Vancutsem, C., Herold, M.; Sambale, J.; Achard, F.; Durieux, L.; Plummer, S. & Weber, J.-L. (2007). GlobCover: ESA Service for Global Land Cover from MERIS, Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2007), IEEE, pp 2412-2415, ISBN 978-1-42441211-2, Barcelona, July 2007

Badea, A.; Dana, I.F.; Moise, C.; Mamulea, A.A.; Olteanu, V.G. & Poenaru, V.D. (2008). Land Cover GIS Databases, Support for the Implementation of the European Agriculture and Environment Programs (2008). 00470048, Annals of DAAAM for 2008 & Proceedings of the 19th International DAAAM Symposium, ISBN 978-3901509-68-1, ISSN 1726-9679, pp 024, Editor B. Katalinic, Published by DAAAM International, Vienna, Austria 2008

Di Gregorio, A. & Jansen, L. (2000). Land Cover Classification System (LCCS): Classification Concepts and User Manual, FAO, ISBN 92-5-104216-0, Italy

Lambin, E. & Geist, H. (2006). Land Use and Land Cover Change--Local Processes and Global Impact, Springer, ISBN: 978-3-540-32201-6, Berlin, Heidelberg

***http://dataservice.eea.europa.eu/dataservice/metadetails.asp ?id=1007 Accesed on: 2009-02-20