Wilderness degradation happens when new encroachments are made changing the wilderness status of an area. It is a complex issue which does not easily lend it self to a GIS based analysis. I will refer to my posting on wilderness for a peek into the complex world of wilderness philosophy.

It is possible to set up a system like FME to do an analysis of changes in wilderness due to new encroachments. The procedure I made generates a wilderness degradation data set based on wilderness and (new) encroachment data. It is based on procedures used by the Norwegian government in their analysis of wilderness and encroachment. The system should however be easy to accommodate for different preconditions by manipulating the number of wilderness zones and/or their buffer distances.

To visualize the degradation of wilderness it is necessary to make a categorization and furthermore establish a cartography to carry the information to the reader. In my view this can not be done unless the author/mapmaker to some extent takes side in what is good or not good related to wilderness degradation. This will be the focus on a forthcoming posting, but I will touch into the issues here as well.

FME is but one of the potential solutions for producing the results. QGIS, GeoKettle and even PostGIS could be good alternatives. The GitHub project has room for alternative implementations.

Before we start – encroachment is the process of making areas accessible. It need not be due to some bad intent. But access to areas means that the areas are subject to any person and their intent. The wilderness degradation assumes a worst-case scenario.

Wilderness categories

Changes in wilderness status can be visualized according to the “loss” of wilderness status for the areas. Or put in an other way – the increased access to areas. In the analysis and method discussed earlier we are using the following categories:

• w wilderness zone more than 5 kms from encroachment
• 1 wilderness zone 3-5 kms from encroachment
• 2 wilderness zone 1-3 kms from encroachment
• n near encroachment 0-1 kms
• e encroachment (in FME this is the “null” category)

The categories and method is based on wilderness analysis as done in Norway over the last 20 years (NEA 2014). Several evaluations has been made (NEA 2014 and ??). Being a technical paper I will not discuss the issues discussed in the mentioned sources.

Wilderness degradation – the categories and cartography

Any given area with one of these categories can transit to an other category. Based on the categories the following transitions are theoretically possible. For all practical purposes it seems degradation is the rule rather than the exception.

In the table above all degradations are illustrated with a red color. No-change-data has blue and all upgraded areas area illustrated with green. The choice if colour is not neutral. If any of you wondered, the author thinks degradation is bad and upgrading of areas good. The latter is more theoretical and will not be part of the analysis – this time.

Not all degradations are aequal. The cartography will reflect this giving the absolute loss of area categories the following colours:

This is of course just one way of choosing a color scheme. An other one could be to consider the relative degradation indicating that a degradation from wilderness to near encroachment has a higher severity than wilderness to zone 1. We could call this relative loss of wilderness.

Based on the above we end up with the following version of the transition table.

Using FME to analyse – the code…

Creating an analysis for degradation is also rather straightforward. Again I wish to do this using open source software, but for now you will just have to deal with FME.

The code making the analysis possible is available here:

• Wildernesss/fme @ github

A practical example

In the example below I will be using wilderness data from an area in Bulgaria and a fictive road intersecting with an existing wilderness area.

As you can see the area starts out with a core of wilderness area.

Next comes a road which connects two areas already subject to encroachment.

If we calculate the areas degraded due to the new road this is what we get.

The new situation can be visualized as such.

The procedure can handle several input objects. If you start out with a country level infrastructure data set you can use the procedure presented in Wilderness 2: A practical approach to wilderness analysis using FME to produce the baseline wilderness data. Using  the difference geoprocessing tool in QGIS (or similar in ArcGIS) on an updated version of the infrastructure data could you could effectively establish a new encroachment data set. Sampling the infrastructure data (from OpenStreetMap or other sources) at regular intervals you could then produce yearly or even monthly wilderness degradation data sets.

Big data

The process could be automated using a dedicated computer. For countries or areas with mature OpenStreetMap data sets this could effectively represent authoritative wilderness maps. For countries with immature data the analysis will at best show where  data is missing.

Assuming that we have access to good data sets for infrastructure the calculations using FME or other processing tools will be demanding.

The procedures presented are based on vector analysis. Doing the procedure using raster analysis is likely to lessen the computation burden. For bigger areas with good data sets the calculations will still run for days.

Concluding remarks

The described method in this posting along with my earlier postings can serve as a starting point for governments or NGOs wanting to establish wilderness data sets and wilderness degradation data sets. The method is suitable as a starting point for discussions about land use in general.

I will conclude this series with a posting on how wilderness and wilderness loss can be visualized. Wont be ready until the beginning of February though.