2003•MAUSAMOpen access

Verification of model predicted precipitation over India during the summer monsoon of 1997

B. K. Basu

Open full text 5 citations

Abstract

For the monsoon season of 1997 the precipitation observation from 1333 rain gauge stations are used to form spatial averages for 135 grid points of the NCMRWF forecast model covering the whole of India. These spatial averages are compared with the model forecasts of precipitation by computing the mean error, mean absolute error, root mean square error and correlation coefficient. The skill of forecast is also computed by discretizing the precipitation amounts in to standard categories. The seasonal accumulated precipitation and its spatial distribution are well reproduced in the model predictions up to 4-days in advance. The mean error in prediction shows spread of precipitation across the Western Ghat hills as the barrier effect is under estimated in the model due to poor horizontal resolution. The correlation coefficient shows good in-phase relationship between the predicted and observed precipitation especially over areas of very large seasonal precipitation. Over a small area this coefficient has values higher than 0.6. The model has skill for categories of precipitation with class mark up to 1.0 inch and prediction lead-time up to 2 day. In 1997 no low frequency south to north propagating mode is present in the observed box average precipitation.

Open-access reader

About this research paper

What this paper is about

For the monsoon season of 1997 the precipitation observation from 1333 rain gauge stations are used to form spatial averages for 135 grid points of the NCMRWF forecast model covering the whole of India. These spatial averages are compared with the model forecasts of precipitation by computing the mean error, mean absolute error, root mean square error and correlation coefficient. The skill of forecast is also computed by discretizing the precipitation amounts in to standard categories. The seasonal accumulated precipitation and its spatial distribution are well reproduced in the model predictions up to 4-days in advance. The mean error in prediction shows spread of precipitation across the Western Ghat hills as the barrier effect is under estimated in the model due to poor horizontal resolution. The correlation coefficient shows good in-phase relationship between the predicted and observed precipitation especially over areas of very large seasonal precipitation. Over a small area this coefficient has values higher than 0.6. The model has skill for categories of precipitation with class mark up to 1.0 inch and prediction lead-time up to 2 day. In 1997 no low frequency south to north propagating mode is present in the observed box average precipitation.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

For the monsoon season of 1997 the precipitation observation from 1333 rain gauge stations are used to form spatial averages for 135 grid points of the NCMRWF forecast model covering the whole of India. These spatial averages are compared with the model forecasts of precipitation by computing the mean error, mean absolute error, root mean square error and correlation coefficient. The skill of forecast is also computed by discretizing the precipitation amounts in to standard categories. The seasonal accumulated precipitation and its spatial distribution are well reproduced in the model predictions up to 4-days in advance. The mean error in prediction shows spread of precipitation across the Western Ghat hills as the barrier effect is under estimated in the model due to poor horizontal resolution. The correlation coefficient shows good in-phase relationship between the predicted and observed precipitation especially over areas of very large seasonal precipitation. Over a small area this coefficient has values higher than 0.6. The model has skill for categories of precipitation with class mark up to 1.0 inch and prediction lead-time up to 2 day. In 1997 no low frequency south to north propagating mode is present in the observed box average precipitation.

Key concepts: Precipitation, Correlation coefficient, Climatology, Environmental science, Rain gauge, Monsoon, Quantitative precipitation forecast, Mean squared error

Related papers

Back to paper searchBrowse research topicsOriginal source
Verification of model predicted precipitation over India during the summer monsoon of 1997 — Research Paper | ScholarLens