Evaluation of ERA5 Global Model Performance in Regional Waters Based on Year-round Meteorological Observations: A Case Study in The Bay of Bengal

Abstract

Monitoring the meteorological parameters is crucial in understanding the day to day weather for planning various activities. Longterm monitoring is essential to estimate the climate change and associated impacts. One of the key adverse effects of climate change on mankind is the increasing number of cyclones of high intensity. Indian moored buoy network in the North Indian Ocean has been maintained by NIOT for the past 2.5 decades providing crucial meteorological and oceanic observations at strategic locations. One of the important parameter is the Sea Level Pressure(SLP) which is the key indicator of the occurrence of any cyclones. Another parameter is the wind speed (U10) which defines the sea-state or the category of the cyclones. Year-round continuous 3 hourly observations of SLP, U10 as well as Air Temperature at a location on the NW Bay of Bengal in deep waters have been utilized. The buoy has captured the cyclones during the year 2005 and some of the deep depressions happened in its neighbourhood. Further, the global model ERA5 data sets have been compared with measurements to assess ERA5 performance in the Indian waters. ERA5 performed satisfactorily with correlation of 99% for SLP, 91% for Air temperatue and reasonably good correlation with 89% and 81% for Wind speed and Direction respectively. The corresponding RMSE are 0.99 ^oC, 0.57 hPa, 1.67 m/s respectively. Monthly statistics have been provided to ascertain the ERA5 performance during different seasons. The performance during the cyclones (PYARR, BAAZ, and FANOOS) and Depressions in close proximity (BOB 02, BOB 04 and BOB 08) is evaluated. From this study it is found that, the ERA5 data sets of Air Temperature and SLP can be used with more confidence than the ERA5 winds and is more so, during high winds of short duration such as cyclones as well as during onset of SW and NE monsoons in May and October, where the discrepancies are more. These findings are in line with the earlier studies.