Pan-tropical Runoff for the World Water Development Report II

Gridded fields of changes in runoff and river flow (discharge) due to
historical and hypothetical deforestation from Douglas et
al. (2005). Scenario 1 (Sc1, historical) compared distributed runoff
(RO) and river flow (discharge, Q) generated from pre-industrial land
cover (based on WWF Terrestrial Ecoregions of the World (Olson et al.,
2001) with runoff and river flow derived from contemporary
landcover. Scenario 2 (Sc2, hypothetical future) compared distributed
runoff and river flow (discharge, Q) generated from the contemporary
landcover with runoff and river flow generated from a hypothetical
future forest conversion scenario.

The contemporary landcover was developed by combining the Global Land
Cover Characteristics Database (GLCCD 2000; Loveland et al. 2000)
interpreted using the International Geosphere-Biosphere Programme
(IGBP) land cover classification scheme (IGBP 1998) with cropland and
cropland mosaic classes from the IFPRI Agricultural Extent database,
which is a reinterpretation of the GLCCD v2.0 dataset. This exercise
identified areas, at a 1km resolution, that contain 30 percent or more
agricultural activity. For details on the creation of the IFPRI
Agricultural Extent, see Wood et al. (2000) and IFPRI (2002). The
Center for Sustainability and the Global Environment at the University
of Wisconsin has developed a global pasture surface based also on a
reinterpretation of the GLCCD v2.0 data. The pasture surface
represents non-forest areas that are used for grazing (Ramankutty
2003). The agricultural extent and pasture surfaces were superimposed
on the IGBP classified GLCCD to create the contemporary land cover
surface.

The objective Scenario 2 (hypothetical future forest conversion), was
to design a hypothetical, "worst case" land cover change experiment
that explored deforestation in the most vulnerable tracts of remaining
forest, and to measure what effect this conversion could potentially
have on biodiversity, hydrological function and ultimately, on
downstream human populations. The baseline for this scenario was the
contemporary land cover surface because the contemporary land cover
surface was our best available representation of existing land
cover. In Scenario 2, only existing forest area (circa 1992/3) within
the tropical forest biomes was targeted for conversion; land cover in
all remaining portions of the domain was held constant. The global
conservation status from the WWF terrestrial ecoregions database
(Olson et al., 2001) was used to identify the areas most vulnerable to
change within the tropical forest biomes; forest conversion was
limited to areas. The conservation status is an indicator of the
degree and threat of change for each individual ecosystem. The key
factors defining the conservation status are the degree of habitat
loss, the level of fragmentation, remaining block size and level of
conversion. The areas classified as critical or endangered were
designated as those most threatened in terms of potential
deforestation. Areas currently under protection, as identified by The
World Conservation Monitoring Centre's (UNEP-WCMC, 2003) protected
areas database, were therefore not subject to conversion.

The sceanrios are defined as:

Scenario 1: Change in runoff (del RO, in mm/yr) due to historical
forest conversion to agriculture: Douglas et al. (2005)

Scenario 1: Change in discharge (del Q, in km3/yr) due to historical
forest conversion to agriculture: Douglas et al. (2005)

Scenario 1: Relative change in discharge (ratio of del Q to Q) due to
historical forest conversion to agriculture: Douglas et al. (2005)

Scenario 2: Change in runoff (del RO, in mm/yr) due to hypothetical
future forest conversion to agriculture: Douglas et al. (2005)

Scenario 2: Change in discharge (del Q, in km3/yr) due to hypothetical
future forest conversion to agriculture: Douglas et al. (2005)

Scenario 2: Relative change in discharge (ratio of del Q to Q)
due to hypothetical future forest conversion to agriculture: Douglas
et al. (2005)