Land cover maps constitute a necessary tool for development planning and management of the territory. Furthermore, land cover maps depicting the current reality are a must in countries where, due to political changes, rapid dynamic phenomena have taken place, resulting in a complete restructuring of the agricultural and other sectors, as in the case of Bulgaria.
The scale of such maps should be large enough to provide detailed information, however it should allow for regional assessment, statistics and subsequent planning. The 1:50 000 scale is the most suited for this exercise.
The project has produced 14 land cover maps at 1:50 000 scale for selected test areas of the country, covering 5 600 sq km. These maps were prepared using Landsat satellite data, acquired in 1998 and 1999 as the main data source and thus represent the land cover existing at that time. The land cover classification was performed using the FAO Land Cover Classification System (LCCS).
Landsat Thematic Mapper and Enhanced Thematic Mapper data, with their very good spectral and spatial characteristics, provide excellent ways for discriminating natural and man-made features and map them at 1:50 000 scale. Subsequent field checking of the interpretation assumptions, performed with the use of GPS, showed a very high accuracy of the mapping.
The classification of the land cover using LCCS in the three test areas resulted in a legend of 49 classes. As the legend was developed in parallel with the classification of the various features using LCCS, it is evident that if this exercise is expanded to cover the whole country, the number of classes will increase. In fact, new land cover classes, typical of particular areas, will have to be added to the general legend. This underlines very well the flexibility and precision of the LCCS methodology, as opposed to land cover classifications using a predefined legend, such as CORINE.
It is interesting to note that Bulgaria is the first European country where land cover mapping using LCCS has been undertaken.
To each mapped unit (polygon), soil type and erosion features were linked as attributes into the GIS system. This created a comprehensive database, which is unique in Bulgaria. The database provides very useful information for agriculture, forestry and urban development planning, for environment protection and for many other applications. The data collected in the database provide the possibility for different kinds of spatial analysis, which is necessary in land management.
Furthermore, the land cover maps and associated database constitute an appropriate basis for the correct application of agricultural statistics. To achieve good results in this field, it is evident that it is important to know the surface devoted to agriculture, its coverage by permanent (orchard, vineyard) and seasonal (cereal, other) crops, the portion not cultivated or used as grassland. As an example of the usefulness of updated land cover maps, the irrigated rice within the study area in the Plovdiv region now (1999) covers only a surface of 743.79 hectares, compared with the thousands of hectares of a few years ago.
Similarly, the same maps and database provide all information on forestry, as classes such as coniferous, deciduous, mixed, riverine forest, forest plantations, wetlands etc, are mapped separately. The same exercise can be undertaken for all land cover features reported in the maps.
As the maps and database are geo-referenced to the national topographical grid, for each mapped unit (polygon) it is possible to know the exact location, its surface coverage, its inherent land cover, the soil type occurring there and the forms of erosion affecting the site.
By adding to the database municipal and district boundaries, it will be possible to extract interesting statistics, such as, for example, agricultural land used in the municipality and so on. To demonstrate this very important aspect of the database, this application was carried out for the village of Rassovo, in the Montana test area, with very interesting results.
By combining the land cover database and the Landsat image with a DEM, the present and potential erosion features of an area in the Sofia region were estimated. This application, now possible in Bulgaria for all the areas covered by the present mapping, provides a very important tool to all involved in land management, mainly in areas bounded by hilly or mountainous terrain.
The project has also used very high resolution IKONOS satellite data, acquired in August 2000, for a variety of applications in the Sandanski area in the southern portion of the Strouma Valley.
The inventory of vineyards is of particular relevance. In the area studied, the data show that vineyards which are no longer cultivated or cut down altogether amount to 35 percent of the total area they were covering previously.
The same satellite data were used for updating the linear features (roads, channels, rivers) of a detailed topographic map at 1:5 000 scale and for improving a detailed soil map at the same scale.
The flexibility and accuracy of the satellite data are evident in all the above cases. However, the very high spatial resolution of the IKONOS data (1 m), resulting in too much detail, and its inherent cost, suggest its use when small areas should be studied in good detail or when the large scale of work required cannot be met by other more economic ways.
Considering the methodological approach and the results obtained by the project and reported in this study, the following main conclusions can be made:
- Landsat ETM satellite data are the most suitable for land cover mapping at 1:50 000 scale. Further, their cost is extremely low. Repetitive coverage and data quality are very good;Based on the above considerations, it is highly recommended to expand the exercise to the whole country, by applying the same methodological approach. Total cost will be very low, approximately US$ 350 000.
- LCCS, now used worldwide, permits land cover classification easily and with accuracy, as it is based on a set of independent diagnostic criteria;
- the database allows for the extraction of a variety of information for agricultural development planning, environment protection and statistics for many applications.
By adding a new layer to the database, that is the municipal and district boundaries, the database will permit any kind of statistics at national, district and municipal level, a powerful tool for inventory, monitoring and decision-making.
As the database has been developed using ArcView, a very common GIS software, it will be easy to combine the database with other data sets, existing or in preparation, for a variety of different applications.
The database could be also used for computerized decision-support systems using application-specific models, for instance to calculate erosion hazard risks or changes in water runoff associated with land cover changes.