5:15 PM - 6:45 PM
[AGE28-P06] Effects of Changes in Rainfall Characteristics due to Climate Change on Erosion Prediction by the Water Erosion Prediction Project (WEPP)
Keywords:Soil erosion, Climate change, WEPP, d4PDF, Short-term rainfall intensity
1. Introduction
Soil erosion significantly contributes to soil degradation of agricultural lands. Billions of tons of soil are lost annually worldwide due to water erosion by rainfall (FAO, 1996). Climate change is expected to exacerbate this issue through more frequent intense and short rainfall events. The Water Erosion Prediction Project (WEPP) employs the CLIGEN weather generator and the MarkSim weather model to simulate future climate data for erosion prediction, incorporating the effects of climate change (Trotochaud et al., 2014). A strong correlation between water erosion and short-term rainfall intensity -peak rainfall values over minutes or hours- has been reported (Nagasawa et al., 1993). However, research on how short-term rainfall characteristics affect WEPP's predictions under climate change is limited. In this study, we clarified the impact of statistical values related to short-term rainfall intensity on the prediction values generated by WEPP by using the hourly weather forecast model, d4PDF (Mizuta et al., 2017).
2. Methods
1) Sensitivity Analysis of CLIGEN and WEPP to Rainfall Parameters
WEPP requires input parameters of soil, topography, plants, and meteorological conditions. This study sought inputs from observed data in a sugarcane field in Ishigaki Island, Okinawa Prefecture, as a baseline, comparing with the study by Machida et al. (2020). A sensitivity analysis was conducted on rainfall parameters significantly contributing to erosion prediction: MEAN P (the monthly average of daily rainfall) and MX.5P (the monthly average of the maximum 30-minute rainfall intensity).
2) Future Erosion Prediction Using Climate Models
Monthly meteorological statistics for erosion prediction were generated using actual measurements, MarkSim, and d4PDF, and simulations for past (2000-2020) and future (2050-2070, 2070-2090) water erosion were conducted.
3. Results
1) Sensitivity Analysis of CLIGEN and WEPP to Rainfall Parameters
The sensitivity analysis showed a higher sensitivity for MX.5P than MEAN P, especially when MX.5P was smaller than 30 mm h-1.
2) Future Erosion Prediction using Climate Models
The results showed that, for the historical period, average annual water erosion predictions made using d4PDF had smaller gaps than those made with MarkSim compared to predictions based on actual measurements. Furthermore, for future predictions, the increase rate of the predicted values using d4PDF was larger than that with the MarkSim. It may be due to the increased predictions of MX.5P, to which WEPP's erosion prediction sensitivity is high.
Soil erosion significantly contributes to soil degradation of agricultural lands. Billions of tons of soil are lost annually worldwide due to water erosion by rainfall (FAO, 1996). Climate change is expected to exacerbate this issue through more frequent intense and short rainfall events. The Water Erosion Prediction Project (WEPP) employs the CLIGEN weather generator and the MarkSim weather model to simulate future climate data for erosion prediction, incorporating the effects of climate change (Trotochaud et al., 2014). A strong correlation between water erosion and short-term rainfall intensity -peak rainfall values over minutes or hours- has been reported (Nagasawa et al., 1993). However, research on how short-term rainfall characteristics affect WEPP's predictions under climate change is limited. In this study, we clarified the impact of statistical values related to short-term rainfall intensity on the prediction values generated by WEPP by using the hourly weather forecast model, d4PDF (Mizuta et al., 2017).
2. Methods
1) Sensitivity Analysis of CLIGEN and WEPP to Rainfall Parameters
WEPP requires input parameters of soil, topography, plants, and meteorological conditions. This study sought inputs from observed data in a sugarcane field in Ishigaki Island, Okinawa Prefecture, as a baseline, comparing with the study by Machida et al. (2020). A sensitivity analysis was conducted on rainfall parameters significantly contributing to erosion prediction: MEAN P (the monthly average of daily rainfall) and MX.5P (the monthly average of the maximum 30-minute rainfall intensity).
2) Future Erosion Prediction Using Climate Models
Monthly meteorological statistics for erosion prediction were generated using actual measurements, MarkSim, and d4PDF, and simulations for past (2000-2020) and future (2050-2070, 2070-2090) water erosion were conducted.
3. Results
1) Sensitivity Analysis of CLIGEN and WEPP to Rainfall Parameters
The sensitivity analysis showed a higher sensitivity for MX.5P than MEAN P, especially when MX.5P was smaller than 30 mm h-1.
2) Future Erosion Prediction using Climate Models
The results showed that, for the historical period, average annual water erosion predictions made using d4PDF had smaller gaps than those made with MarkSim compared to predictions based on actual measurements. Furthermore, for future predictions, the increase rate of the predicted values using d4PDF was larger than that with the MarkSim. It may be due to the increased predictions of MX.5P, to which WEPP's erosion prediction sensitivity is high.