Using MODIS imagery to map cultivated areas in West Africa
The northern fringe of sub-Saharan Africa is a region that is considered particularly vulnerable to climate variability and change, and where food security remains a major challenge. One of the preliminary stages necessary for analyzing impacts of West African Monsoon and its variability on agriculture and food security is a reliable estimation of the cultivated domain at national level, a scale compatible with climate change studies. The opportunity of using satellite remote sensing for agricultural statistics has been explored by the research community as well as by national departments of agriculture during the last few decades [1]. In Africa, existing global land cover maps result from different initiatives such as the GLC2000 or the POSTEL global land cover maps [2] but generally they are more focused on ecosystems than on agricultural systems. In the Sub-Saharan Africa countries, operational land cover mapping schemes are restricted by the cost of high resolution images. Yet, the monitoring of vast ecosystems at national or continental scale typically resorts to low-resolution free images [e.g. 3], but the pixel size of these images is generally too coarse for the identification of fields, especially in fragmented landscapes. Nevertheless, recent moderate-resolution sensors, such as MODIS/TERRA, with spatial resolutions as low as 250 m, offer new possibilities in the study of agricultural lands. With this increased spatial resolution, the detection of groups of fields can now be considered. The low and medium spatial resolutions do not, by themselves, provide a completely satisfactory representation of the landscape but are compensated for by a large coverage area and an excellent temporal resolution. This brings about the question whether moderate-resolution satellite data, in combination with external data (thematic maps, statistics, etc.) can provide a correct assessment of the distribution of the cultivated domain at country level. It is expected that more consistent information on vegetation would allow to monitor Sahelian rural landscapes with better continuity, thereby providing, combined with crop models, relevant information for early warning system.
| Main Authors: | , , , , |
|---|---|
| Format: | conference_item biblioteca |
| Language: | eng |
| Published: |
Agropolis international
|
| Subjects: | U30 - Méthodes de recherche, F01 - Culture des plantes, B10 - Géographie, |
| Online Access: | http://agritrop.cirad.fr/557266/ http://agritrop.cirad.fr/557266/1/ID557232.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The northern fringe of sub-Saharan Africa is a region that is considered particularly vulnerable to climate variability and change, and where food security remains a major challenge. One of the preliminary stages necessary for analyzing impacts of West African Monsoon and its variability on agriculture and food security is a reliable estimation of the cultivated domain at national level, a scale compatible with climate change studies. The opportunity of using satellite remote sensing for agricultural statistics has been explored by the research community as well as by national departments of agriculture during the last few decades [1]. In Africa, existing global land cover maps result from different initiatives such as the GLC2000 or the POSTEL global land cover maps [2] but generally they are more focused on ecosystems than on agricultural systems. In the Sub-Saharan Africa countries, operational land cover mapping schemes are restricted by the cost of high resolution images. Yet, the monitoring of vast ecosystems at national or continental scale typically resorts to low-resolution free images [e.g. 3], but the pixel size of these images is generally too coarse for the identification of fields, especially in fragmented landscapes. Nevertheless, recent moderate-resolution sensors, such as MODIS/TERRA, with spatial resolutions as low as 250 m, offer new possibilities in the study of agricultural lands. With this increased spatial resolution, the detection of groups of fields can now be considered. The low and medium spatial resolutions do not, by themselves, provide a completely satisfactory representation of the landscape but are compensated for by a large coverage area and an excellent temporal resolution. This brings about the question whether moderate-resolution satellite data, in combination with external data (thematic maps, statistics, etc.) can provide a correct assessment of the distribution of the cultivated domain at country level. It is expected that more consistent information on vegetation would allow to monitor Sahelian rural landscapes with better continuity, thereby providing, combined with crop models, relevant information for early warning system. |
|---|