Authors: Junyu Lu*, University of South Carolina, Gregory J. Carbone, University of South Carolina
Topics: Climatology and Meteorology, Hazards, Risks, and Disasters, Soils
Keywords: Agricultural drought, Climate projection uncertainty, Signal to noise ratio, CMIP5 multi-model ensembles
Session Type: Paper
Start / End Time: 2:40 PM / 4:20 PM
Room: Jackson, Marriott, 5th Floor
Presentation File: No File Uploaded
Future climate changes are very likely to alter hydrometeorological patterns and change the characteristic of drought from regional to global scale. However, there are still considerable uncertainties in drought projections. Here, we focus on agricultural drought and use surface soil moisture output from CMIP5 multi-model ensembles under RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. We investigate the seasonal and annual percentage change of surface soil moisture and evaluate the statistical significance of change using student t-tests for each grid. The annual mean soil moisture by the end of 21st century shows statistically significant large scale drying and limited areas with wetter conditions for all scenarios, with stronger drying as the strength of radiative forcing increases. The median frequency of short-term drought (shorter than 6-months) is projected to decrease, while the median frequency of long-term drought (longer than or equal to 6-months) is projected to increase. Further, we quantify and partition three sources of uncertainty: internal variability, model uncertainty, and scenario uncertainty in the projection of agricultural drought. Variability between models presents the largest source of uncertainty (over 80%) over the entire 21st century owing to the simplified hydrological models in many CMIP5 climate models and complicated process controlling soil moisture. Finally, we also examine the spatiotemporal variability of annual and seasonal signal to noise (S/N) change across the global and for different lead times. The spatial pattern and magnitude of S/N do not change significantly by lead time, indicating that the spread of uncertainties become larger as the signals become stronger.