Introduction | Methodology | Results | Conclusions | Recommendations

LUCID Projects - Uganda Sites: Methodology

In each of the three sites, quantitative land cover/use inventory was carried. This was done for 1955 for all the three sites, 1984 for Sango Bay and LMNP and 2000 for all the three sites. The land cover/use for each site (Sango Bay, LMNP & Kabale/Ntungamo Border) for 1955, 1984 and 2000 were produced as described below:

First, extraction of accurate and consistent spatial information from affordable remotely sensed data for savannah and small-scale farmed ecosystems that account about 80% of Uganda’s landscape is not a trivial matter. First, standard automated image classification techniques are based on a cartographic paradigm whose foundation is visual image classification. It is of practical importance to use image resolution (scale) as a guiding principle for defining an appropriate cartographic scale (and minimum mapping unit) when extracting information from remotely sensed data. However, it is questionable whether the cartographic paradigm holds if medium and coarse imagery are used for mapping savannah and small-scale farmed ecosystems. Nonetheless, medium to coarse resolution imagery, combined with field observations, are indispensable techniques for generating spatial land cover information over large areas.

Second, for a long time, it has been assumed that use of medium to coarse resolution imagery for mapping heterogeneous ecosystems (such as savannah and small-scale farming) is justified because imaging sensors such as Landsat and AVHRR “automatically aggregate the spectra of heterogeneous into homogeneous land surface features. This assumption appears not to be true. This is because, according to Mugisha (2002), spectra of heterogeneous landscapes such as savannah ecosystems aggravate image misclassification irrespective of the cartographic scale selected for mapping.

What was needed was an appropriate framework for land cover mapping and monitoring using medium to coarse resolution imagery in Uganda’s dominant ecosystems. A method developed for generating consistent, accurate and affordable spatial information from Landsat ETM+ imagery for both savannah and small-scale farmed ecosystems was developed during this study for Uganda’s ecosystems. The method is based on the premise that Landsat M/ETM+ data have sufficient information to allow updating detailed spatial information generated from high resolution imagery (both panchromatic and multispectral) even for savannah and small-scale farmed ecosystems that dominate Uganda’s landscapes. For further details of the method developed for monitoring and mapping land cover/use in Uganda’s savannah and small-scale farmed ecosystems, refer to Uganda LUCID Working Paper Number one.

The method developed as described in Uganda LUCID Working Paper Number One was the basis of carrying out land cover/use change analysis in three Uganda LUCID study sites ( Sango Bay, LMNP and Kabale/Ntungamo Border ), and three other sites for but related projects. The status of land cover/use for Kibale National Park and adjacent areas are included in this publication. Results of quantitative land cover/use change analysis for each of the four sites are summarised Table 1 . The aim of carrying out land cover/use change analysis was to find out the root causes of land cover/use changes in Uganda over the past 100 years. Both quantitative and qualitative data were collected and analysed to identify the major root causes of land cover/use in Uganda. The major findings of the study can be summarised as follows:

Using a combination of both qualitative and spatially explicit data, It was been possible to demonstrate that land cover/use changes are triggered by specific events and then accelerated by a tandem of many biophysical-socio-economic factors in Uganda. The current land use patterns in Uganda may be attributed to many driving forces including droughts, presence or absence of disease, civil strife, cultural impacts and government policies that over the past 100 years. The study concludes that knowledge of the root causes of land cover/use changes in a given ecosystem provides valuable insights into other key environmental and socio-economic processes such as biodiversity loss, land degradation, societal conflicts and food insecurity. Understanding the linkages between land cover/use change and key biophysical-socio-economic processes is also beneficial to the formulation of policies for sustainable use of Uganda's natural resources base.

For details of the land cover/use change analysis for each of the three sites, the reader is referred to Uganda LUCID Working Paper Number Two.