BNPP/ASB Functional Value of Biodiversity Project – Phase II 

i Assemble more detailed information on biodiversity-rich tropical habitats (IFPRI lead initiative)

iv Undertake synoptic modeling of hydrological impacts of land use change

 1Aiii. Task 2c. Create the “current” land cover surface

The ‘current’ land cover surface is a compilation of 2 different interpretations of the GLCCD v2.0 dataset, the University of Kassel’s irrigated area dataset (GMIAv2.1), Nightlights of the World (ORNL), and SAGE’s pasture land surface.  Detailed documentation on the steps taken to create the current surface are included in current_lc.doc.  The basic steps are as follows:

• Create a revised classification of GLCCD based on a combination of classes from IGBP & IFPRI’s agricultural extent surface

• Identify urban areas based on the CityLights database using a 50 percent threshold

• Calculate the shares of each land cover class within each unique 0.5dd cell of the pantropic unique id surface

• Disaggregate the area under mosaic classes into individual classes for each cell (e.g. break Agriculture / Forest mosaic into 50 percent agriculture and 50 percent forest)

• Calculate the area within each cell that is equipped for irrigation

• Adjust the existing land cover classes to incorporate the irrigated area by first taking the amount of irrigated area from the agricultural area; if irrigated area exceeds agricultural area then adjust the remaining land cover classes equally to result in the same total area for that cell

• Calculate the share of each cell occupied by each land cover class and determine the aggregate land cover value for each cell

• Review methodology with UNH / make adjustments if necessary / provide UNH with final surface and documentation

• Pantropic Basin extent (UNH 2003)

• GLCCDv2.0 (land cover data) – IGBP classification (USGS/EDC)

• Agricultural extent based on a reinterpretation of GLCCDv2.0 (IFPRI)

• CityLights (ORNL)

• GMIA v2.1 – University of Kassel  percent Irrigated Database

• Pastureland surface – University of Wisconsin, Center for Sustainability and the Global Environment (SAGE)

General notes, 

Comments 

The rules by which contemporary data on the extent of converted land cover classes (e.g. cropland, pasture and urban) are assumed to have been converted from “pre-industrial” land cover classes. Complexities arise since all pixel scales used – including the 1 km resolution source data – contain mosaics within which constituent land classes can be identified, at best, by area shares rather than by their sub-pixel location.

An important goal was to find some way to retain Pan-Tropic “Scenario 1” – assessing the scale of habitat/biodiversity loss and changes in hydrological regime associated with conversion of our target tropical forest biomes from pre-industrial to contemporary times. Given the problems being experienced in reconciling various sources of land cover information to develop this Scenario as well as in obtaining intuitive/defensible hydrological impacts, a view was expressed that we could postpone analysis of this historic perspective and focus only on the forward looking scenarios using contemporary land cover as a basis. However, we are very reluctant to shortcut this scenario since it provides (or not) a significant proportion of our credibility base regarding the data and modeling approaches we are employing. Getting the wrinkles ironed out of this scenario will, we believe, both facilitate and significantly improve the reliability of results obtained from the forward-looking scenarios (as well provide a yardstick for the scale of impacts they are likely to generate).

Recapping Scenario 1 (‘contemporary’). To minimize problems of reconciling natural ecosystem classes between the pre-industrial (WWF ecoregion) and contemporary (IBGP) map legends, Scenario 1 has been defined as examining the difference between

The pre-industrial (WWF ecoregion) land cover, and

The pre-industrial land cover superimposed with converted (agriculture and urban) areas within the 3 focus tropical forest biomes

In the case of Scenario 1, the way in which information on agricultural land cover shares are netted out of (converted from) pre-industrial land cover class shares was reviewed. It was felt that the rules currently being applied were biasing conversion away from forested areas (and hence, reducing the average hydrological signal associated with conversion). It was agreed that IFPRI would sample the highest resolution (1km) land cover data by region in order to derive regional “rules of thumb” for the relative contribution of natural vegetation classes to agricultural land at the scale required by the hydrological model (50km). This analysis also feeds into item 2 above. Kate will supply the new “current” landcover surface along with the landcover conversion rules to Ellen so that it can be converted to WBM classes[1].

                                                                                                                                               back to top

Global Partnership Amazon Congo Basin Southeast Asia
Climate Change Biodiversity Sustainability Poverty, Policy and Deforestation
Partners Publications Policy Briefs Gallery Funding Priorities Get Involved
e-mail: asb@cgiar.org   Home Page

Last updated: 26 November, 2003     ©2003 ASB. All rights reserved.