Conclusion

In summary, we offer the human modification framework as an explicit approach to better quantify the spatial patterning and degree to which locations have been altered by human activities. We found that in our case study area, HMF showed markedly different results, particularly in the spatial patterns and distribution of disturbed areas, as compared to GAP status levels. An important, nonintuitive result is that we found that level 4 had less evidence of alteration than did level 3. This can occur because the HMF relies on surrogate variables that characterize current land use activities that have been shown to have effects on biodiversity. In contrast, GAP status levels characterize the legal protection from land use conversion (or management plans)—not the current condition per se. It is important to note that the HMF, as employed in this paper, shows existing (or at least currently known) modifications to the landscape, whereas the GAP and IUCN approach estimate long-term protection (in perpetuity). Although we did not document these trends here, the HMF methods could be easily applied to future growth or potential land use scenarios to examine the differences between their effects. The fine-grained approach and spatial context analysis in particular would be sensitive to differences in patterns of potential future land uses.

We believe that current approaches to estimating the degree of human modification can be improved in three ways. First, as understanding and elicitation of threats to protected areas has matured, we believe that approaches will need to move from using general surrogates to factors that are more closely linked to the threats, ideally in a mechanistic way. In addition, we believe that finer-grained data are increasingly required to address concerns about fragmentation and other pattern issues. Finally, the spatial (or perhaps landscape) context is critical to incorporate. Further work is needed to develop methods for generating landscape context that more directly reflect ecological processes. For example, neighborhoods for freshwater should use a series of flow-connected watersheds. Similarly, the flow or movement of processes through other transport vectors such as air, soil transport, animal movement, seed dispersal, etc., should be explicitly represented as well.

There are a number of data gaps that need to be filled. Some of these gaps could be filled by collecting data from field offices, standardizing the data and entering them into an electronic database (with spatial coordinates recorded). For example, grazing permit data have been collected for many decades, but much of it remains in an unstandardized, unmapped, and nonelectronic format. The location of timber harvests (year, harvest method) data are available, but not in a consistent, easy-to-use format. Another type of data gap would simply require acquisition and compilation of existing spatial databases, (e.g., highway structures such as fencing, culverts, bridges, etc., or hiking trails that would include structures, tread type, width, use, etc.). Spatial data about water infrastructure, in addition to large dams, would be useful as well for structures such as ditches, levees, rip-rap, and channelized segments.