Research prototype weather pattern forecasts for India

12 UTC GEFS-IMD medium-range run on Mon 27 Mar 2023
Forecasts valid at 12:00 UTC / 17:30 IST

Weather pattern forecast probabilities

Ensemble members are objectively assigned to their closest matching weather pattern definition, using a set of 30 predefined weather patterns. The assignment is achieved by finding the ensemble member and weather pattern pairing with the smallest area-weighted sum of squared differences between their corresponding u- and v-component winds at 850-hPa – this metric is commonly referred to as "distance". Forecast probabilities (as shown in the table below) are then based on the total number of ensemble members assigned to each weather pattern. Click on the "Pattern" links in the first column to view weather pattern definition wind fields at 850-hPa as well as their rainfall climatologies. Click on the "Historical frequency occurrences" links in the last three columns to view a full breakdown for all months in the year. Table cells are colour-coded according to their probability to assist with identifying the most likely weather pattern transitions.

 
 Tue 
 28 
 Mar 
 Wed 
 29 
 Mar 
 Thu 
 30 
 Mar 
 Fri 
 31 
 Mar 
 Sat 
 1 
 Apr 
 Sun 
 2 
 Apr 
 Mon 
 3 
 Apr 
 Tue 
 4 
 Apr 
 Wed 
 5 
 Apr 
 Thu 
 6 
 Apr 
 Pattern Category Historical frequency
occurrences
(February)		 Historical frequency
occurrences
(March)		 Historical frequency
occurrences
(April)
Pattern 1 Retreating Monsoon 1.2% 0.8% 0.2%
Pattern 2 5% Spring Dry Period 11.0% 4.8% 0.5%
Pattern 3 Spring Dry Period 10.0% 4.4% 0.4%
Pattern 4 Break Monsoon 0.0% 0.0% 0.0%
Pattern 5 57% 33% 5% 86% 86% 52% 24% 19% Western Disturbances 3.0% 13.7% 31.5%
Pattern 6 Retreating Monsoon 0.8% 0.8% 1.2%
Pattern 7 90% 38% 52% 10% 14% 5% 5% 14% 67% Spring Dry Period 14.8% 11.5% 7.5%
Pattern 8 10% 5% 5% Spring Dry Period 12.8% 19.8% 4.8%
Pattern 9 14% Spring Dry Period 10.8% 3.7% 0.4%
Pattern 10 Break Monsoon 0.0% 0.0% 0.0%
Pattern 11 Break Monsoon 0.0% 0.0% 0.0%
Pattern 12 Pre-Monsoon 0.9% 6.1% 21.8%
Pattern 13 Pre/Post-Monsoon 0.0% 0.0% 1.6%
Pattern 14 Pre/Post-Monsoon 0.0% 0.0% 0.0%
Pattern 15 Pre/Post-Monsoon 0.0% 0.2% 1.3%
Pattern 16 Spring Dry Period 9.4% 13.1% 2.0%
Pattern 17 Active Monsoon 0.0% 0.0% 0.0%
Pattern 18 Active Bay of Bengal 0.0% 0.1% 0.2%
Pattern 19 Active Monsoon 0.0% 0.0% 0.0%
Pattern 20 Spring Dry Period 11.3% 6.6% 1.7%
Pattern 21 Active Monsoon 0.0% 0.0% 0.0%
Pattern 22 Pre/Post-Monsoon 0.0% 0.0% 0.2%
Pattern 23 14% 14% 43% 67% 38% Western Disturbances 2.9% 7.7% 9.8%
Pattern 24 Western Disturbances 0.1% 0.8% 11.0%
Pattern 25 Active Bay of Bengal 0.0% 0.1% 2.1%
Pattern 26 Monsoon Onset 0.0% 0.0% 0.0%
Pattern 27 14% 71% 5% 29% 10% Western Disturbances 11.0% 5.7% 1.8%
Pattern 28 Active Bay of Bengal 0.0% 0.0% 0.0%
Pattern 29 Retreating Monsoon 0.0% 0.2% 0.0%
Pattern 30 Active Bay of Bengal 0.0% 0.0% 0.0%

Cumulative probabilities for each weather pattern category

Probabilities are aggregated for weather patterns with the same descriptions in the table above. For example, probabilities for all weather patterns described as "active monsoon" will be aggregated. This helps identify transitions in the large-scale circulation and is especially useful when there is a lot of ensemble spread across the set of 30 weather patterns.

 
 Tue 
 28 
 Mar 
 Wed 
 29 
 Mar 
 Thu 
 30 
 Mar 
 Fri 
 31 
 Mar 
 Sat 
 1 
 Apr 
 Sun 
 2 
 Apr 
 Mon 
 3 
 Apr 
 Tue 
 4 
 Apr 
 Wed 
 5 
 Apr 
 Thu 
 6 
 Apr 
Pre-Monsoon
Pre/Post-Monsoon
Active Monsoon
Western Disturbances 57% 48% 90% 86% 100% 95% 95% 86% 10%
Monsoon Onset
Spring Dry Period 100% 43% 52% 10% 14% 5% 5% 14% 90%
Break Monsoon
Active Bay of Bengal
Retreating Monsoon

Circulation trend indicator tables

Weather patterns have been classified according to their dominant circulation characteristics during four broad seasons: (1) summer monsoon, (2) summer to retreating monsoon, (3) winter dry period and (4) pre-monsoon to monsoon onset. Circulation trends are shown in the four tables below for both the latest and most recent forecasts with daily validity times aligned. This allows for a rapid assessment of forecast consistency over consecutive model runs. In addition, these tables are useful for identifying key transitions in the large-scale circulation.

1. Summer monsoon circulation trend

Please note: This table is intended for use between June and September for identifying active and break phases of the summer monsoon.

 
Tue
28
Mar
Wed
29
Mar
Thu
30
Mar
Fri
31
Mar
Sat
1
Apr
Sun
2
Apr
Mon
3
Apr
Tue
4
Apr
Wed
5
Apr
Thu
6
Apr
Latest model run (27/12 UTC) 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%
Previous model run (run -1) 100% 100% 100% 100% 100% 100% 100% 100% 100%  
Previous model run (run -2) 100% 100% 100% 100% 100% 100% 100% 100% 100%  
Previous model run (run -3) 100% 100% 100% 100% 100% 100% 100% 100%    
Previous model run (run -4) 100% 100% 100% 100% 100% 100% 100% 100%    
Previous model run (run -5) 100% 100% 100% 100% 100% 100% 100%      
Previous model run (run -6) 100% 100% 100% 100% 100% 100% 100%      
Previous model run (run -7) 100% 100% 100% 100% 100% 100%        
Previous model run (run -8)                    
Previous model run (run -9)                    

            
 Weather patterns typically associated with break monsoon tendencies over India are most likely 
 Weather patterns 4, 10 and 11 (break)
            
 Weather patterns typically associated with no active or break monsoon tendencies over India are most likely 
 Weather patterns not listed above and below
            
 Weather patterns typically associated with active monsoon tendencies over India are most likely 
 Weather patterns 17, 19 and 21 (active)

2. Summer to retreating monsoon circulation trend

Please note: This table is intended for use between August and November for identifying the start of the retreating monsoon.

 
Tue
28
Mar
Wed
29
Mar
Thu
30
Mar
Fri
31
Mar
Sat
1
Apr
Sun
2
Apr
Mon
3
Apr
Tue
4
Apr
Wed
5
Apr
Thu
6
Apr
Latest model run (27/12 UTC) 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%
Previous model run (run -1) 100% 100% 100% 100% 100% 100% 100% 100% 95%  
Previous model run (run -2) 100% 100% 100% 100% 100% 100% 86% 100% 100%  
Previous model run (run -3) 100% 100% 100% 100% 100% 81% 86% 90%    
Previous model run (run -4) 100% 100% 100% 100% 100% 100% 100% 90%    
Previous model run (run -5) 100% 100% 100% 100% 100% 100% 95%      
Previous model run (run -6) 100% 100% 100% 100% 100% 100% 100%      
Previous model run (run -7) 100% 100% 100% 100% 100% 95%        
Previous model run (run -8)                    
Previous model run (run -9)                    

            
 Weather patterns typically associated with summer monsoon tendencies over India are most likely 
 Weather patterns 12, 13, 15, 15 and 22 (pre/post summer monsoon); 26 (onset); 17, 19 and 21 (active); 4, 10 and 11 (break) 
            
 Weather patterns typically associated with no summer or retreating monsoon tendencies over India are most likely 
 Weather patterns not listed above and below
            
 Weather patterns typically associated with retreating monsoon tendencies over India are most likely 
 Weather patterns 1, 6, 18, 25, 28, 29 and 30 (retreating)

3. Winter circulation trend

Please note: This table is intended for use between December and May for identifying periods most susceptible to western disturbances during the winter dry period.

 
Tue
28
Mar
Wed
29
Mar
Thu
30
Mar
Fri
31
Mar
Sat
1
Apr
Sun
2
Apr
Mon
3
Apr
Tue
4
Apr
Wed
5
Apr
Thu
6
Apr
Latest model run (27/12 UTC) 100% 57% 52% 90% 86% 100% 95% 95% 86% 90%
Previous model run (run -1) 86% 100% 100% 71% 100% 100% 100% 90% 67%  
Previous model run (run -2) 100% 90% 71% 86% 81% 100% 86% 100% 90%  
Previous model run (run -3) 81% 100% 95% 86% 100% 81% 86% 86%    
Previous model run (run -4) 100% 100% 90% 81% 76% 100% 100% 67%    
Previous model run (run -5) 100% 81% 90% 71% 90% 100% 90%      
Previous model run (run -6) 100% 100% 57% 86% 76% 81% 100%      
Previous model run (run -7) 100% 100% 95% 95% 67% 76%        
Previous model run (run -8)                    
Previous model run (run -9)                    

            
 Weather patterns typically associated with contrywide dry period tendencies over India are most likely 
 Weather patterns 2, 3, 7, 8, 9, 16 and 20 (countrywide winter dry period)
            
 Weather patterns typically associated with no dry period or western disturbance tendencies over India are most likely 
 Weather patterns not listed above and below
            
 Weather patterns typically associated with western disturbance tendencies over India are most likely 
 Weather patterns 5, 23, 24 and 27 (western disturbances)

4. Spring to summer monsoon circulation trend

Please note: This table is intended for use between May and July for identifying monsoon onset.

 
Tue
28
Mar
Wed
29
Mar
Thu
30
Mar
Fri
31
Mar
Sat
1
Apr
Sun
2
Apr
Mon
3
Apr
Tue
4
Apr
Wed
5
Apr
Thu
6
Apr
Latest model run (27/12 UTC) 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%
Previous model run (run -1) 100% 100% 100% 100% 100% 100% 100% 100% 95%  
Previous model run (run -2) 100% 100% 100% 100% 100% 100% 86% 100% 100%  
Previous model run (run -3) 100% 100% 100% 100% 100% 81% 86% 90%    
Previous model run (run -4) 100% 100% 100% 100% 100% 100% 100% 90%    
Previous model run (run -5) 100% 100% 100% 100% 100% 100% 95%      
Previous model run (run -6) 100% 100% 100% 100% 100% 100% 100%      
Previous model run (run -7) 100% 100% 100% 100% 100% 95%        
Previous model run (run -8)                    
Previous model run (run -9)                    

            
 Weather patterns typically associated with pre-monsoon tendencies over India are most likely 
 Weather patterns 12, 13, 14, 15 and 22 (pre-monsoon)
            
 Weather patterns typically associated with no pre- or summer monsoon tendencies over India are most likely 
 Weather patterns not listed above and below
            
 Weather patterns typically associated with summer monsoon tendencies over India are most likely 
 Weather patterns 26 (onset); 17, 19 and 21 (active); 4, 10 and 11 (break)

Weather regime stacked probabilities

Weather pattern probabilities are aggregated according to their common regime group and presented as stacked probabilities. Regime bars are colour-coded based on their dominant circulation characteristics. Here, orange indicates winter types, light blue indicates pre- and post-monsoon / monsoon onset types, dark blue indicates the main summer monsoon types (active / break) and green indicates the retreating monsoon type. Hatching is used to distinguish between sub-types within a given colour. The regime ordering always remains the same to allow for an easy comparison with forecasts from previous runs and from different models. These plots complement the circulation trend indicator tables by enabling a quick assessment of any large-scale circulation changes and can be compared with previous runs to interpret forecast consistency. Regime descriptions in the plot are for guidance only; please refer to the top-level weather pattern forecasts for more detail.

Forecast confidence index

The forecast confidence index plot allows for an objective assessment of ensemble spread across the set of 30 weather patterns, by comparing the forecast confidence index from the latest run to that from a number of historical runs (as defined in the plot title). A score of 0 means that all members are equally distributed across all weather patterns, whereas a score of 1 means that all members are assigned to the same weather pattern. Green shading highlights days where ensemble spread is less than normal and red shading highlights days where ensemble spread is more than normal. Note that the forecast confidence index is not a measure of forecast skill.

Underlying forecast data (CSV files)

CSV files are available for download which provide the underlying data used in the generation of these forecasts. This includes the weather pattern assignments for each ensemble member and the distances (as a grid-point average area-weighted sum of squared differences) between each ensemble member and its closest matching weather pattern definition, using u- and v-component winds at 850-hPa between both fields. CSV files contain one row for each forecast day. The first column gives the forecast validity date and time in the format YYYYMMDDHH. The following columns then give the values (pattern assignments or distances) for each ensemble member, with the control member values given first followed in order by all other members.

Ensemble member pattern assignments
(CSV file)

Ensemble member pattern distances
(CSV file)

Page created at 01:22:30 (India local time) on Tue 28 Mar 2023

Data provided by the India Meteorological Department, MoES, GoI  Colloboration with Met Office  Under WCSSP Project


© Crown copyright 2023