Assessment of the weather research and forecasting model generalized parameterization schemes for advancement of precipitation forecasting in monsoon‐driven river basins

Abstract

Some of the world's largest and flood‐prone river basins experience a seasonal flood regime driven by the monsoon weather system. Highly populated river basins with extensive rain‐fed agricultural productivity such as the Ganges, Indus, Brahmaputra, Irrawaddy, and Mekong are examples of monsoon‐driven river basins. It is therefore appropriate to investigate how precipitation forecasts from numerical models can advance flood forecasting in these basins. In this study, the Weather Research and Forecasting model was used to evaluate downscaling of coarse‐resolution global precipitation forecasts from a numerical weather prediction model. Sensitivity studies were conducted using the TOPSIS analysis to identify the likely best set of microphysics and cumulus parameterization schemes, and spatial resolution from a total set of 15 combinations. This identified best set can pinpoint specific parameterizations needing further development to advance flood forecasting in monsoon‐dominated regimes. It was found that the Betts‐Miller‐Janjic cumulus parameterization scheme with WRF Single‐Moment 5‐class, WRF Single‐Moment 6‐class, and Thompson microphysics schemes exhibited the most skill in the Ganges‐Brahmaputra‐Meghna basins. Finer spatial resolution (3 km) without cumulus parameterization schemes did not yield significant improvements. The short‐listed set of the likely best microphysics‐cumulus parameterization configurations was found to also hold true for the Indus basin. The lesson learned from this study is that a common set of model parameterization and spatial resolution exists for monsoon‐driven seasonal flood regimes at least in South Asian river basins.

Assessment of the Weather Research and Forecasting (WRF) model for simulation of extreme rainfall events in the upper Ganga Basin

“…Overall, the choice of CU appears to have a significant impact on the simulation of rainfall over the region. This conclusion is consistent with the earlier studies such as those by Sikder and Hossain (2016), where they found ISMR to be more sensitive to CU than to MP.…”

“…Some possible factors that could contribute would be that local boundary formulation in MYJ may be more appropriately capturing the vertical environment in the complex terrain compared to the non-local YSU scheme which seeks to simulate vertical mixing and boundary layer evolution using averaged and grid representative fields. With regard to the BMJ CU emerging in the top configuration, there are a number of studies for the ISMR where it has emerged as performing "overall best" (Vaidya and Singh, 2000;Ratnam and Kumar, 2005;Vaidya, 2006;Rao et al, 2007;Kumar et al, 2010;Mukhopadhyay et al, 2010;Srinivas et al, 2013;Sikder and Hossain, 2016). As for the MP scheme, there are limited studies in comparison to those that have studied the CU configuration for the ISMR.…”

“…However, there is no updraft and downdraft of mass flux and no cloud detrainment. Domain 2b is configured without any CU scheme, assuming MP can explicitly solve the convection at the finer resolution (Sikder and Hossain, 2016). The four MP schemes considered are the Purdue Lin (PLin) scheme (Lin et al, 1983;Chen and Sun, 2002), the Eta Ferrier (Eta) scheme (NOAA, 2001), the WRF Single-Moment 6-class (WSM6) scheme (Hong and Lim, 2006) and the Goddard scheme (Tao et al, 1989).…”

Sensitivity Study on High-Resolution WRF Precipitation Forecast for a Heavy Rainfall Event

“…When the horizontal resolution is in the <10 km range, the precipitation behavior of a numerical model with and without those parameterized schemes should be explored systematically. Increasing the skill of forecasting precipitation with various parameterized schemes remains for further study [32].…”

Comparing Machine Learning and Decision Making Approaches to Forecast Long Lead Monthly Rainfall: The City of Vancouver, Canada

“…The second approach is based on TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution), a multi-criteria decision making method, which tests different combinations of predictors, among those identified by MRMR, and investigates all the possible sets of data for rainfall modeling. TOPSIS has been more employed in social studies and for decision making (e.g., [33][34][35]), however, is has recently been shown to be powerful in hydrologic and weather forecasting as well [16,36]. Finally, the ANN and TOPSIS models with the best forecasting performance and their corresponding sets of predictors are identified and compared to propose a promising approach for long lead monthly rainfall forecasting.…”