Publication Date:
2023-06-13
Description:
In this study we are focusing on the retrieval of quantitative precipitation estimates using C-band dual polarization radar installed at southern tip of India. Knowledge of spatial variability of precipitation would be helpful to better understand the dynamics of mesoscale cloud systems. A popular approach is to use doppler weather radar (DWR) data to obtain precipitation using some relationships. Usually power-law type relationships (Z = aR〈sup〉b〈/sup〉) between the rain rate (R) reflectivity (Z) are widely used for retrieval of rainfall. But Z-R relations have a great degree of uncertainty because of the large scatter in the Z-R scatter diagram which is usually used to obtain the Z-R relation. To overcome this issue, we have tried other relations between rain rate and polarimetric variables [e.g., Z〈sub〉h〈/sub〉, Z〈sub〉dr〈/sub〉, rho, K〈sub〉dp〈/sub〉]. we have adopted the T-matrix formulation to calculate the values of Z〈sub〉h〈/sub〉, Z〈sub〉dr〈/sub〉, rho, K〈sub〉dp〈/sub〉 for different 1-minute DSDs observed from optical disdrometer during pre-monsoon [Mar-Apr-May] of 2016, 2017, 2019, 2020, 2021. Then we have obtained relations between rain rate and different combinations of polarimetric variables [e.g., R(Z〈sub〉h〈/sub〉), R(K〈sub〉dp〈/sub〉), R(Z〈sub〉h〈/sub〉, Z〈sub〉dr〈/sub〉), R(K〈sub〉dp〈/sub〉, Z〈sub〉dr〈/sub〉)]. Further we evaluated them using the data of 2018 from the polarimetric DWR to find the best estimate among them. We have also tried a machine learning (ML) model to obtain precipitation estimate. Though the overall performance of different estimates is close to each other, there is hint that the ML model might be more useful in heavy precipitation.
Language:
English
Type:
info:eu-repo/semantics/conferenceObject
Permalink