Project Description | Algorithms | Verification | Data | Updates | Evaluations
If you are not a forecaster in either HPC or SPC and you are using this information, please let us know. Thank you!
ATTENTION: The
data provided on this web site is experimental.
Permission is not granted to redistribute
this data
or to put a link to these data from any web page except for NOAA internal use.
The following calendars for November '01 through April '02 allow you to select the daily matrix or individual day products listed below:
|
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The metafile collection should have a default image grouping. If not, please use Arrange and click on Auto-group to get the image groups listed for selection.
The precipitation type determination from each algorithm individually is presented in the "algrtyps" metafiles. The results are displayed at 3-hour intervals and should be considered to apply instantaneously at the posted time. Please click on the examples below:
The most probable precipitation type is that type that has the highest relative frequency (with the aforementioned weighting) among the algorithms at each point at each hour. The result is displayed as a weather symbol at each point. The metafile ("probptyp") contains a map displaying the likely type for each hour of the forecast. At intervals (6-hours) the model predicted precipitation contours are overlayed on this display. The labeling shows the precipitation accumulation period. An example for a single hour is shown in here:
A second metafile grouping shows the most probable precipitation type from the specified model with the AVN predicted precipitation overlain. The model precipitation type is instantaneous at the valid time. The AVN forecast precipitation accumulation period is indicated in the label (6 hours). Here is an example from this group:
The time averaged and interval maximum values of the probability of freezing rain, snow, ice pellets, and rain are displayed in the "risks" metafile. The intervals, in forecast hours, are 6-18, 12-24, 18-30, 24-36, and 30-48. The accumulated precipitation contours for these periods are overlain. The probabilities are displayed as color coded symbols, where green indicates low risk (10-30%), yellow moderate risk (31-60%), and red high risk (>60%). These probabilities are, again, the relative frequencies of occurrence of the given precipitation type in the ensemble of algorithms. In this case, the values have been averaged over time, or selected over time for the maxima. Examples for some of these forecast time ranges may be selected below:
Ntrans will automatically group the time intervals.
Precipitation type guidance is computed for the 00Z and 12Z cycles of the RUC and Eta models. The source of data for the calculations is hourly surface and profile data from each model in a BUFR file. The BUFR data are first reformatted into GEMPAK sounding and surface data files. Then a collection (ensemble) of precipitation type determination algorithms assembled by John Cortinas (CIMMS) is executed. This yields an ensemble of precpitation type determinations at each point for each hour of the model forecast, since the point data is dumped hourly. The point locations scattered across the CONUS roughly correspond to hourly reporting stations with the addition of TAF sites in the case of the Eta. The distribution and density of the stations will be apparent when viewing the displays.
Three Ntrans metafiles are generated for each cycle of each model. The names of these files will appear among the meta file names listed when you select "eta" or "ruc" in the "Select Model" button board display in Ntrans. The generic names of these files are listed below, where YYMMDD is the year, month, and day date, HH is the cycle hour, and MMM is the model:
More information about the content of these files is given below, with examples of displays. Until they appear in Ntrans, these files can be found by using the User-Defined option in the Ntrans file selection drop-down menu.
For more information about which GEMPAK variables were used to create each metafile, click here.
The probability of a given precipitation type is its relative frequency of occurrence in the ensemble of algorithms. The algorithms are invoked only if the model hourly precipitation accumulation exceeds .1 mm. So, if the precipitation threshold is not reached, the probabilities are all zero. Therefore, non-zero probabilities are conditional on the existence of precipitation in the model.
John Cortinas has done some verification studies that suggest that some algorithms are better than others for a given precipitation type. Using this information, the best algorithm for each type listed above is given a weight of 2.0. The following table gives the name of the algorithm that has a double weight for each precipitation type.
| TYPE | ALGORITHM |
|---|---|
| Freezing Rain | RAMER |
| Snow | Czys |
| Ice Pellets | NCEP |
| Rain | RAMER |
| Mixed Precip | RAMER |
| Indeterminate type | RAMER |
Note that the RAMER indeterminate type is also double weighted to assure that probabilities sum to 1 for the collection of ensembles. There is no assurance from data that the RAMER code (or any code) is best at judging an indeterminate situation. Fortunately, this situation does not appear to occur frequently.
