Geographic information systems (GIS) tool are becoming increasingly important in conservation and natural resource management planning and implementation. The good news is that it is also getting easier to find relevant and freely available (spatial) data sets online. However, we also need to be careful in evaluating the accuracy and reliability of these data sets, as illustrated by an article I came across today.
In the article Implications of Spatial Data Variations for Protected Areas Management: An Example from East Africa the authors Dowhaniuk, Hartter and Ryan (2014) compare three different spatial boundary data of protected areas for seven Ugandan parks. The data sets they compare are from the International Union for the Conservation of Nature, World Resource Institute, and Uganda Wildlife Authority. Their results show discrepancies of 0.4% to 9.0% between these three data sets. The authors argue that this may lead to erroneous findings in e.g., larger-scale assessments of deforestation or evaluations of the effectiveness of local-scale management for encroachment, thus emphasizing that in a world were conservation or land use managers increasingly rely on spatial data sets, data quality and accuracy is a key issue.
The possible discrepancies between mapped and actual on-the-ground boundaries is also illustrated by a new layer that I added to the online potential natural vegetation map for eastern Africa. It shows the potential natural vegetation in the nationally designated protected areas only. The map was created by clipping the potential natural vegetation map using the Protected areas from the World Database on Protected Areas (WDPA). The resulting map provides an easy one-click overview of where what vegetation types are (nominally) protected.
Zooming in on some of the parks will show you that for some parks there is a clear discrepancies between the mapped boundaries and what appears to be sharp boundaries in vegetation cover. Take for example the map of Unyambiu North, a small protected area in Tanzania. You can see a clear differences in vegetation cover within and outside the protected area (image below). This sharp boundaries could suggest that land use pressure in this part of Tanzania is high, and that the protected status is effective in preventing the conversion of the natural vegetation into agricultural areas. What it also shows, however, is that the boundaries of the park WDPA and the actual boundaries on the ground are misaligned.
This example (and you can find many more if you browse the map) thus illustrates that we really need to be careful and take into account potential errors in our spatial data sets. Of course, the comparison of the protected areas databases is just one example. Another good example is the discrepancy in land cover classification between different remote-sensing based land cover maps (see e.g., appendix 2 in environmental gap analysis to prioritize conservation efforts in eastern Africa).
To see the layer I am talking about, first open the online map: online vegetation map. Next, click on the little icon in the lower right corner (the layer icon).
In the popup menu, select vegetation in ppa layer and deselect the layer Vegetation. In the example below I selected the satellite layer to compare actual vegetation boundaries with boundaries of the protected areas.