2007年-世界发展银行全球_The_Impacts_of_Climate_Change_on_Regional_Water_Resources_and_Agriculture_in_Africa_68页_2mb

Summary of "The Impacts of Climate Change on Regional Water Resources and Agriculture in Africa"

Core Content

This working paper presents a hydrological assessment of the potential impacts of climate change on water resources and agriculture in Africa, with a focus on the use of the WatBal (Water Balance) model. The study is part of a larger project examining the sensitivity of agro-ecological systems to climate change and the economic implications of adaptation strategies. The findings are intended to be used in Ricardian regressions, which assess the impact of climate variables on land value and farm revenue.

Main Objectives

1. Develop a method to simulate changes in soil moisture and runoff across the continent.
2. Generate time series data (1961–1990) for runoff and flow to serve as a baseline for climate change scenarios.
3. Update hydroclimatic data to the year 2000 using new input data and scenarios.
4. Provide climate change scenario analyses for use in agricultural economic assessments.

Key Findings

• The WatBal model is used to simulate runoff and soil moisture changes at a 0.5° latitude/longitude grid across Africa.
• The model integrates climate data (1961–1990) and physiological parameters (soil properties, land use) derived from global datasets.
• The model outputs include:
  • Monthly time series of simulated runoff
  • Relative soil moisture storage (z, 0–1)
  • Potential and actual evapotranspiration
  • Temperature, precipitation, and streamflow
• The study focuses on 11 countries, representing a wide range of climatic conditions.
• Districts are the primary unit for analysis, and their boundaries are based on geographical data from various sources, including USAID, FEWS, and the FAO geonetwork.
• Irrigated area and river density index are calculated as indicators of surface water availability and potential for irrigation development.
• The model was calibrated using UNH runoff data and WBM simulated data due to limitations in observed flow data.
• A genetic mutation algorithm was employed to optimize model parameters (α, ε, AWC_mult) to minimize root mean square error and improve the model's accuracy in simulating runoff volumes.
• The model was applied to simulate flow across the continent using the STN-30p drainage network, which routes runoff to the ocean.
• The results were used to estimate flow values for each district, based on the weighted average of runoff from contributing grid cells.

Methodology

• The WatBal model is a conceptual rainfall-runoff model that simulates changes in soil moisture and runoff.
• The model uses three key parameters:
  • α (sub-surface runoff coefficient): Influences sub-surface runoff and is higher in regions with greater runoff.
  • ε (surface runoff exponent): Determines the non-linear relationship between surface runoff and storage.
  • AWC_mult (maximum rooting depth): Related to land use and vegetation type, influencing surface runoff.
• The model is applied to a 0.5° grid covering the entire continent.
• Runoff simulation for the period 1961–1990 is conducted, and results are aggregated to the district level.
• Flow simulation is based on a drainage network, with the STN-30p model used to simulate the direction and accumulation of runoff.
• The river density index is calculated to reflect the frequency of streams and surface water availability within each district.

Key Information

• Model inputs:
  • Climate data from the Climate Research Unit (CRU).
  • Physiological data on soil water holding capacity and land use from FAO and IIASA.
  • Drainage network data from the University of New Hampshire (UNH).
• Calibration:
  • Based on UNH simulated runoff and WBM data.
  • Parameters are optimized using a genetic mutation algorithm.
  • AWC_mult is constrained between 0.1 m and 5.0 m, depending on land use class.
• Data processing:
  • Runoff data is converted to district level using geoprocessing tools in ArcView.
  • River density index is calculated by converting district area to square kilometers and summing river lengths within each district.
• Climate change scenarios:
  • 16 scenarios are generated using four climate models (CSIRO2, HadCM3, CGCM2, ECHAM and PCM) and two emission scenarios (A2 & B2).
  • These scenarios are used to simulate the potential impacts of climate change on runoff and evaporation.

Conclusion

The study provides a foundational hydrological assessment for understanding the impacts of climate change on water resources and agriculture in Africa. It emphasizes the importance of hydrological modeling and Ricardian regression analysis in evaluating climate change effects on land value and farm revenue. The model's results are essential for informing future adaptation strategies and policy decisions in the region.