Background Information On Summer Rainfall Predictability Over Texas
- The forecast is generated using the methodology described in Fernando, D.N., Chakraborty, S., Fu, R. and Mace, R.E., 2019. A process-based statistical seasonal prediction of May–July rainfall anomalies over Texas and the Southern Great Plains of the United States. Climate Services, 16, p.100133 .
- The forecast tool is built on an understanding of the spring drivers of intense summer drought over Texas.
How To Use And Interpret The Seasonal Rainfall Forecasts
- The forecast maps provide information on whether the average rainfall for the May‒July (MJJ) season will be above-normal, near-normal, or below normal in the MJJ season, where “normal” refers to the average 3-month rainfall observed for MJJ over the time period 1982‒2016.
- The forecasts for the MJJ season are made from January through end-April. The forecasts are updated at a bi-weekly (15 day) interval. Therefore, there are eight forecasts in all (mid-January, end-January, mid-February, end-February, mid-March, end-March, mid-April, and end-April).
- All but the mid- and end-April forecasts of MJJ rainfall are based on predicted values of select April atmospheric circulation patterns and soil moisture over Texas, which are known to be influential in driving MJJ rainfall. The mid- and end-April forecasts are based on actual observations of these atmospheric circulation patterns and soil moisture over Texas.
- The forecast is expressed as probabilities for each of the three possible categories — below-normal, near-normal, or above-normal — occurring.
- If there is no clear signal of whether the coming season is going to be wetter or dryer than normal, each category would have an equal likelihood of occurring. Each category would then be assigned a 33.3% probability of occurrence.
- If, however, there is a clear signal that the coming season is going to be wetter or dryer than normal, then either the wetter or the dryer category will have a greater probability of occurring. For example, if there is a clear signal that the coming MJJ season is going to be dryer than normal, the probabilities for each forecast category might be 55% for the below-normal category, 20% for the near-normal category, and 25% for the above-normal category.
- The forecast provides information on the likelihood of above-, near-, or below-normal MJJ rainfall for each county in Texas.
- The forecast does not provide information on how much wetter or dryer than normal each county is going to be in the MJJ season. The forecast only provides an estimate of what the chances are for each county to experience an MJJ season that will be dryer or wetter than the long-term average.
- Red color on the map indicates chances for below-normal (dry) conditions, blue color indicates chances for above-normal (wet) conditions, and green color indicates chances for near-normal (average) conditions.
- For each county, only the highest probability value will be shown in shading on the map.
- Hovering over a county with the mouse opens up a popup, which provides the probabilities associated with each category for the selected county.
Probability of Exceedance Graphs
- The pop-up graph, known as the Probability of Exceedance Graph, provides information on the point forecast (or predicted average 3-month rainfall) for each county.
- The stepped blue curve shows the exceedance probabilities associated with different values of observed rainfall over the MJJ season. The smooth orange curve shows a log-normal fit to the observed rainfall exceedances. These curves convey information on the probability that a given quantity of rainfall will be exceeded. Lower amounts of rainfall (left end of the x-axis) have higher probabilities of exceedance and higher amounts of rainfall (right end of the x-axis) have lower probabilities exceedance.
- The Probability of Exceedance graphs can be used to determine the probability that the quantity of rainfall in the coming season will fall within an upper and lower limit or below a certain value (e.g. 10th percentile of climatological rainfall for the season).
- The green dot indicates the current forecast rainfall value. Hovering the cursor over the green dot provides information on the forecast rainfall amount, the probability that this value will be exceeded [referred to as Exceedance (%)] on the plot], and the upper and lower limit of the forecast. The upper and lower limit of the point forecast is based on a confidence interval of 60%. This information is provided to convey the uncertainty associated with the point forecast.
- This page shows the predictions for MJJ of the current year at the beginning of May of the current year using observations from April of the current year. It also shows the predictions that were issued versus the observed rainfall for MJJ for all years going back to 2015, which was the first time we issued an actual forecast.
- The 2010‒2014 hindcasts show what the statistical model generated as the forecast for MJJ 2010, MJJ 2011, MJJ 2012, MJJ 2013, and MJJ 2014 based on April observations from 2010, 2011, 2012, 2013, and 2014, respectively. The observed rainfall for the MJJ season in these five years is also shown.
What about rainfall forecasts for other seasons besides May‒July?
- The rainfall forecasts provided are only for the MJJ season because the tool is specifically designed to incorporate physical mechanisms in the spring known to influence summer rainfall over Texas.
- Further research on drivers of rainfall in other seasons is needed before a tool to forecast rainfall in these seasons can be developed.
Limitations, Data Needs, and Potential Opportunities
- This rainfall forecast is based on current understanding of factors driving summer rainfall over Texas. As research on this subject continues, and new knowledge on factors driving summer rainfall is generated, we will make efforts to enhance the tool.
- The accuracy of datasets used as predictor variables, particularly the soil moisture input dataset, would influence the accuracy of the forecast. Improved soil moisture observations, whether from in-situ stations or from remotely sensed data products, could potentially enhance the accuracy of the forecast.
If you have questions or need further information, please contact us.
The provision of these rainfall forecasts via this website was made possible through the U.S. Bureau of Reclamation WaterSMART Drought Resiliency funding to the Texas Water Development Board (Grant: R15AP00184; Fiscal Year: 2015).