This is a marker for work in progress. Do not cite.
This morning I was contacted in order to provide geographical information for one of our partners in the Reforlan project. The aim of the research is to find priority areas for forest restoration.
The question of how to go about the analysis is thought provoking. I have already mentioned the problem of aggregation error in this weblog. Pooling data in order to find mean values does not work at the scale in which most of the important processes act. The classic example is mean rainfall for Mexico. At the same time, some elements on the landscape cannot be mapped at a fine scale, and others, such as management units (farms or ejidos) form natural aggregation according to social factors.
So perhaps the first step would be determine priority pixels according to a limited number of physical processes that can be resonably accurately mapped at fine resolution (30 m x 30 m) then aggregate to find the number of pixels in each agronomic unit that could be restored. An analysis of social and economic factors at this scale could then be undertaken in order to discover where restoration might be feasible and bring the most social benefit.
The mappable factors at the 30m pixel scale would be
1. Forest cover and recent deforestation: Pixels that have forest cover presumably do not need restoring. Pixels that have been recently deforested may be prioritised, although natural regneration may be a suitable mechanism in some areas close to the original forest.
2. Slope: Areas with steep slopes must be prioritised for restoration in order to avoid soil erosion. Factors such as slope length may be taken into account.
3. Insolation (aspect): South facing slopes that receive intense insolation during the dry winter months may be in need of restoration, but successful restoration may be more challenging.
4. Distance from rivers: Riparian areas may be in greater need of restoration and trees may establish more successfully in moist soil. At the same time these may be the most valuable areas for irrigated agriculture.
5. Conservation status: Reserves should perhaps take priority when restoration is being considered.
The interesting element of the research involves finding a suitable way of combing the layers in order to allow ranks to be calculated. A typical approach is weighted averaging, but this may not allow interactive effects to be included.
The map below shows the estimated current extent of the forest in the region and recent deforestation hotspots.
Areas with the steepest slopes do tend to coincide with remaining forest cover, as would be expected, but some recent deforestation both intentionally and as a result of forest fires has taken place in the steeply sloping “Sierra” region in the West and South West and South of the study region.
North facing slopes receive less direct sunlight during clear winter days and are more likely to be covered in evergreen forest. Fuel dries out on south facing slopes leading to more intense fires. Regeneration may be difficult due to hydric stress.
Most deforestation has occured close to the major rivers.
It may also be useful to evaluate soil type (not shown) although the maps tend to be at a coarse scale.
Once pixels have been ranked according to aptitude and priority for restoration (the two factors may be inversely related in many cases) they may be aggregated to agricultural management units for further analysis of socio-economic factors.