This thesis looks at the capabilities of satellite remote sensing for use in environmental impact assessment (EIA), using an agricultural project area in Southwestern Tanzania as a case study.
The “cloud” based GIS software Google Earth Engine is utilized for the analysis, which consists of four parts: land use mapping, a normalized differentiated vegetation index (NDVI) time series chart, the extent of dry season crops, and differences in NDVI for various part of the agricultural areas.
The results show that 50% of the study area consists of fields and have a very high accuracy for all land cover classes, except for built up areas. The NDVI time series chart illustrates that better cloud filtering should be investigated in order to have a clearer view of the annual phenology. Furthermore, the extent of fields with dry season crops have been measured to be 20%, which is only 1/3 of what other sources state. This indicates that further research is needed to look into this gap. However, the areas found to have dry season crops correspond well with areas that have been mapped to have a high yearly NDVI, which could suggest that these areas might have better growing conditions and/or better farming practices being applied to them.
No field studies have been conducted in the case study area because of logistical and financial constraints, and it is therefore recommended that the findings from this thesis are validated through fields surveys in the area. This approach should always be strived for when conducting remote sensing analysis, including EIA projects.
This thesis also recognizes that remote sensing cannot aid in all issues regarding EIAs in developing countries, as not all are related to a lack of data but instead societal structures such as poverty, lack of well-functioning government institutions, corruption etc. Hence, remote sensing is deemed to be one of many approaches needed for improving EIAs and environmental protection.