About the Data

The maps of the human footprint and of the last of the wild should not be read too literally. Although there is no doubt that the human footprint and last of the wild express an important perspective on the world, it is also true that, in its details, these maps contain inaccuracies (acknowledged below) and are mapped at a scale coarser than most conservation efforts

These maps are based on geographic proxies for drivers of human impact:  human population density, land cover and land use mapping, lights regularly visible from a satellite at night, locations of roads, rivers and coasts, settlement patterns, etc.  However drivers are not inevitably impacts.  One of the roles of conservation is to find ways for human beings to lessen their impacts while still making their living.

The input datasets used to map the human footprint are enormously expensive to maintain and update, as a result they tend to lag behind the patterns they depict. 

All the datasets used here were released in the 1990s, however some are based on much older datasets or datasets which are incomplete. 

 In all cases they do not depict the current 2002 extent of roads, settlement or population density. 

This problem leads to underestimates of the amount of human influence. 

The lights data sometimes over-estimates the “lit” area for over-bright pixels.  The population dataset relies on population estimates made in different countries using different census techniques, which sometimes results in marked changes across national boundaries.  These problems probably lead in most instances to underestimates of human influence, but may result in overestimates in some cases.

Our interpretations of the amount of human influence based on the input factors relies on simplified scales from 0 to 10 which do not vary by region, biological or cultural context.  The understanding of the human influence on nature is in its infancy, and despite 100 years of ecological science, not known very well; however we do know enough to be concerned.  We tried to be conservative and common sensical in our determinations of human influence from the various input factors, using advice from the scientific literature and our colleagues.

We probably overestimate the direct effect of roads in some cases.  Direct influence from roads in terms of pollution, soil compaction, modification of stream courses and waterways, introduction of new species, and road kill is known to vary from a few meters to up to several hundred meters from roads. The roads dataset we used maps roads only to an accuracy of 2 km, so we treated all of this 2 km region as influenced by roads. We also treat human access from roads as up to 15 km from roads of all types, though this may be less in some countries and more in others.

The level of access from rivers is probably also incorrectly estimated in some instances.  We defined access along all major rivers, where a major river was defined as one depicted as one or more polygons in the input database and connected continuously to the sea.However access along all rivers is probably more likely, since any river large enough to be mapped is probably large enough to support a canoe or other boat.  But our rivers dataset does not include the effects of waterfalls or dams, which can impede access up rivers.

The human footprint and the last of the wild do not directly take into account war and conflict between groups of human beings, though these effects may dramatically influence the outcomes of human influence on wildlife and wild places.  In some cases, for example, conflicts lead to increases in the levels of hunting because of increased access to weapons, even long after the war has ended (for example, in Cambodia.)  In other cases, conflicts result in lower human population densities and less investment in infrastructure (for example, in Angola) with the result that areas become wilder.

The human footprint and the last of the wild are not the complete story of conservation.  In fact they do not directly consider conservation targets (animals, plants, air, soil or water processes) at all. Conservation planning requires understanding what is important to conserve in a given area (the conservation targets), how those conservation targets respond to human influence, and the type and degree of that human influence.  The last of the wild is not a complete prescription for nature conservation.  Even if we saved all of the last-of-the-wild areas, our task would be incomplete.

Specific Points:

• Brazil – a tile of the roads dataset is missing in central Brazil along the eastern edge of the Amazon rainforest in the region of the Chapadas das Mangabieras.
• Democratic Republic of Congo – another tile of roads is missing in the central part of the country.
• New Guinea – in Irian Jaya human influence from the Taritatu River is over estimated, because although it is a large river connected to the sea, there is little human movement up the river into the Foja Mountains Reserve, a relatively pristine area that appears heavily influenced on the human footprint map.  In general the human footprint map seems to over-estimate influence in many parts of New Guinea and should be used with caution.

Tundra and boreal forest biomes – Access during winter months is not restricted to roads or rivers, but can occur wherever the snow is packed enough to support a motor vehicle.

Note: We welcome specific comments about the human footprint and the last of the wild, especially notations of specific areas where the level of human influence seems to be over- or under-estimated.  Send your comments to last-of-the-wild@wcs.org.