Along-trajectory rainfall rate

Topics about the HYSPLIT trajectory model.
Post Reply
Posts: 15
Joined: July 1st, 2016, 1:05 pm
Registered HYSPLIT User: No

Along-trajectory rainfall rate

Post by ryanxia » October 27th, 2017, 4:54 pm


I am interested in along-trajectory meteorological parameter: rainfall rate
I displayed GDAS data text profile and found in 2D field of each 3h interval, there is a "TPP6" (mm): I think it is the accumulated precipitation in the last 3h. Is that right? Is this interpolated from station data or forecasted by the GDAS model? How reliable are the rainfall data? Because I want to only analyze rainfall-producing trajectory, I wish I could use this to screen out dry trajectories. I wrote codes to do this. The weather station rain gauge nearby have a lot of daily data gaps...
In addition, after I generated trajectory, there is a column of rainfall rate (mm/hr) data along the trajectory. How does the HYSPLIT calculate rainfall rate? I found rainfall rate data at the starting point doesn't match the 2D field rainfall data. For example, the rainfall rate of the starting point is 0.3 mm/hr, but the 2D field rainfall amount is as high as 5.0 mm.

Another question is about the rainfall data from ERA-interim model. I have no problem in generating ERA-interim reanalysis files including both analysis and forecast fields (I asked question and get answered from this forum in this summer), but I found in 2D field of each 6h interval, there is a "TPPT" (mm). Is it different from "TPP6" in GDAS?

Also, I found GDAS and ERA-interim data don't match very well in terms of rainfall data (and rainfall rate in trajectories.), although all other data, like wind speed, temperature, potentials, and trajectory pathways are similar.

A lot of question... thanks!!


Posts: 254
Joined: November 9th, 2012, 4:23 pm
Registered HYSPLIT User: Yes

Re: Along-trajectory rainfall rate

Post by barbara.stunder » October 31st, 2017, 10:18 am

A description of the 4-character field identifier in the HYSPLIT-formatted meteorology files is at

TPPT is 12-hour accumulation.

TPP6 is 6-hour accumulation. This mean that at hour 03z, it is the accumulated precipitation from 00z to 03z. At hour 06z, it is the accumulation from 00z to 06z. HYSPLIT subtracts the 03z value from the 06z value to get the accumulated precip between 03z and 06z. If you run the HYSPLIT program 'contour' (in the GUI, Meteorology / Display Data / Contour Map) for GDAS, you see the precipitation at 00z, 06z, 12z, and 18z, is accumulated over 6 hours.

This precipitation is the precipitation in the GDAS model. How accurate is it? You can estimate it through a comparison with the nearby rain gauge you mention.

For TPP6, the rainfall rate that is calculated along a trajectory is the 3-hour accumulation at a location divided by 3 hours.

Does this clarify your discrepancy between rainfall rate at the trajectory start point and the 2D dump from the profile program? If not, what is your trajectory start location and date/time?

Probably the precipitation differences between GDAS and ERA-interim just show the difficulties of modeling precipitation.

Post Reply