An analogue method for simultaneous prediction of surface weather parameters at a specific location in the Western Himalaya in India

Abstract Avalanches in the Western Himalaya cause loss of life and damage property every year. To reduce such losses avalanche forecasting is practiced. This technique requires information about various surface weather parameters at least a couple of days in advance. In view of the above requirement, an analogue method has been employed to predict maximum, minimum and ambient temperatures, average wind speed, surface pressure and relative humidity at a specific location in the Western Himalaya. Predictions for surface weather parameters 3 days in advance have been attempted using information of past synoptic situations corresponding to the current situation searched from a database through Euclidean distance. Surface weather observations of the past 16 winters (1989–1990 to 2005–2006) have been used to develop the model. The model has been tested with independent data set of 166 days of winter 2006–2007. The forecast accuracy for maximum, minimum and ambient temperature falls in the range 74.7–85.5, 64.2–81.8 and 58.5–79.9% within an error limit of 3 °C for day 1, day 2 and day 3. Forecast accuracy for prediction of average wind speed is in the range 65.7–70.9% and it is in the range 73.7–82.5% for atmospheric pressure with same error limit as temperature (with respective units) for day 1, day 2 and day 3. Accuracy for the prediction of relative humidity is in the range 58.5–69.3%, within an error limit of 20% for all three days. The root mean square error (RMSE) and mean absolute error (MAE) of each weather parameter have been computed to test the prediction error of the model. The results suggest that the analogue method holds promise for simultaneous prediction of surface weather parameters with reasonable accuracy. Predicted surface weather parameters have been used for assessing future avalanche situations with reasonable confidence. Copyright © 2008 Royal Meteorological Society