Skip navigation
Air Resources Laboratory banner image
Air Resources Laboratory web site National Oceanic and Atmospheric Administration (NOAA)

HYSPLIT Frequently Asked Questions

To find the answer to a question, click on the Click for answer. symbol.

HYSPLIT in General

Meteorological Data

  • Click for answer.What meteorological archive data sets can I use to model Europe with HYSPLIT?
  • Click for answer.What archived data are available on the web to run HYSPLIT with?
  • Click for answer.What data is used to calculate the trajectories and what is the the vertical resolution of the data?
  • Click for answer.I have a question about section five "meteorological input data" in the HYSPLIT_4 USER's GUIDE. In page 49, what are the meanings of "pole latitude & longitude", "reference latitude & longitude", "synch latitude & longitude", "synch point X and Y" and the relation (or difference) between "reference grid size" and "number X & Y points"? Or would you tell me what are their values if I have gridded meteorological data within lat. 5 N to 60 N and long. 110 E to 140 E., one degree resolution and I want to plot the back-trajectory within the same area?
  • Click for answer.I am trying to better understand what the best approach is for calculating the HYSPLIT trajectories I need. I have mostly looked at using the EDAS data, but it has a lot of holes in it for the period I am interested in, so we wind up with trajectories truncated or not calculatable. I tried to fill in the holes with the FNL data. However, in one case when I calculated 120-hour backward trajectories from Potsdam, NY at 1800 on April 8, 1999 using both the EDAS and FNL data bases, I found that the EDAS data suggested the air was in the upper midwest 5 days prior, but the FNL put it into northern Quebec. Why the difference?
  • Click for answer.Where can I find gridded meteorological data to simulate the Chernobyl accident?
  • Click for answer.I have FNL meteorological files, which I use to compute trajectories with HYSPLIT. I need to extract the numerical value of some fields (geopotential at 700 hPa at each i,j grid point). I ran your program example which extracts the first field (precipitation), but I do not know how modify the program to extract geopotential at 700 hPa.
  • Click for answer.I would like to extract surface winds from the archived EDAS data into ESRI Shape "shp" format, MapInfo "mif" format, or something similar. Is this possible?
  • Click for answer.What archived data can I use prior to the start of the FNL dataset?
  • Click for answer.Can I use surface observations in HYSPLIT?
  • Click for answer.How do you parse GRIB files? NCAR/NCEP reanalysis data has all fields for a certain variable for an entire year in one large GRIB file. I'd like to extract (keeping the data in GRIB format) just one or several of the grid numbers from different files into one file containing all data for a given date/time. This is a requirement to run the ARL HYSPLIT converter to get the data in a format compatible with HYSPLIT. Does any one have a GRIB extractor or editor that allows one to parse grib files?
  • Click for answer.I want to extract the u and v fields from the FNL archive according to the pressure level in some locations. Is this possible?
  • Click for answer.Can HYSPLIT read other forms of meteorological data?
  • Click for answer.I will be converting ECMWF grib files into ARL-packed format for use in calculating backward trajectories in Europe using your HYSPLIT model. What fields do I need to have for HYSPLIT to properly calculate my trajectories?
  • Click for answer.How do I convert my ECMWF GRIB data files into ARL format using ecm2arl on the PC?
  • Click for answer.Could you explain the equation (Zagl = ak2 + bk + c) used by HYSPLIT? What is the different between Ztop, Zgl, Zmsl and Zagl?
  • Click for answer.My back trajectory run today looks different than the same trajectory I ran yesterday using the GDAS current7days archive. How often is the GDAS archive current7days file updated?
  • Click for answer.Can I use WRF data with HYSPLIT?
  • Click for answer.What principles are used for computing the mixing depth from the NCAR/NCEP reanalysis data?
  • Click for answer.What is the smallest horizontal resolution allowed by the code, e.g., 10x10 meters, or 1x1 km?

Model Setup & Parameters

  • Click for answer.Can HYSPLIT be setup to run a short-range dispersion calculation? If so, how?
  • Click for answer.Which value should I input as the start height (MAGL)? What does "above model ground level" mean?
  • Click for answer.I computed a vertical sounding at position (40.24, -3.40) with the FNL archived data. In the text results, the PRSS and the MSLP are 923.3 and 1021 mb, respectively. Assuming a hydrostatic atmosphere the model terrain height at this point can be estimated as 977 MASL. But in the text an Estimated Surface Height of 634 MASL is also reported, which is very close to the actual terrain height of 667 MASL. What should I consider as the model terrain height ASL at this point, 977 or 634 MASL?
  • Click for answer.I have a question regarding what start height to make the trajectories and the correction factor for MAGL...
  • Click 
	              for answer. If I choose AGL for the start height, does that mean that for the entire model run the computed heights are relative to the ground? I'm concerned about this because of the terrain in the mountains. I don't want values to be "underground."
  • Click for answer.When I get backward trajectories starting at a rural East Coast site in NJ, which starting height(s) do you suggest is (are) appropriate? The terrain of the site is flat and near sea level. I am also wondering how long of a trajectory duration is recommended.
  • Click for answer.I am attempting to calculate back trajectories for event based precipitation samples to trace sources of atmospheric pollutants (mercury, nitrogen, pesiticides). I am uncertain as to what height or pressure I should use in the model.
  • Click 
              for answer. If I know the in-situ measured SO2 concentration (pptv), corresponding location, altitude (4 km), and time of measurment, is it possible to calculate a back trajectory of that SO2 plume to know about the spreading over that source region?
  • Click for answer.I'm interested in the backward trajectories at the costal site Baia Terra Nova, Antarctica, for aerosol studying. As in most of the Antarctic Plateau there are mountains, and I wonder if this is the reason why the meteorological model used by HYSPLIT "sees" the Antarctic continent so high (even though it is a coastal site)? Can I obtain backward trajectories at a lower height than the model terrain height (say at sea level), and if not, can I be confident in those trajectories starting from a much higher level than the level of the site under study?
  • Click for answer.What would be the most accurate way to determine back trajectories (BT) for airmass sources at a particular field site. The sampling time at the field site was ~24 hours. Should I run a 24 hr BT for every hour during sampling? Would I then combine all of these BT's to get an accurate BT for the airmass collected over the 24 hr sampling period? Also what height would you suggest to end the BT at, 10 m or 950 h Pa? (April 2004)
  • Click for answer.Is terrain included in the HYSPLIT model and how accurate are surface trajectories computed for complex terrain?
  • Click for answer.Is there anyway to change the simulation specification so that the concentration measurements are standardized across both:
    1. One location at different points in time and
    2. Different locations at the same point in time?
  • Click for answer. Is the model designed only to model slow releases of chemicals or could it also be used in the case of explosions?
  • Click for answer. Would HYSPLIT work to model an atmospheric release of a biological toxin?
  • Click for answer. Can I run 1 trajectory every 6 hours every day for an extended period of time?
  • Click for answer. Is there a batch job that can be submitted to the server to obtain 10-d back-trajectories for the time period specified?
  • Click for answer. Is there a way to run multiple trajectories with a single command in hysplit?
  • Click for answer. Is the time used to calculate trajectories UTC or local time?

Interpretation of Trajectory Results

  • Click for answer.After calculating backward isentropic trajectories on the PC and looking at the plot, particularly the vertical motion plot at the bottom, I see motion across theta surfaces. My question is, shouldn't I see vertical motion on a constant theta (potential temperature) surface?
  • Click for answer.After calculating backward isobaric trajectories I see that the pressure is not constant and looking the vertical motion plot at the bottom of the plot, I see vertical motion (about 500m difference from the starting point to some other points ). Why is this?
  • Click for answer.I am trying to do trajectory analysis for an event in which we have observed meteorological data. Is it possible to download HYSPLIT to the PC and feed it observed data from meteorological stations?
  • Click for answer.When I ran a foward trajectory from a location on a previously computed backward trajectory, I did not return to the starting location of my backward trajectory. Why?
  • Click for answer.Could you tell us how the precipitation rates are determined which accompany the air parcel trajectories in the tdump file?
  • Click for answer.Is it possible to overlay several trajectories onto one map?
  • Click for answer.How many trajectories can I draw in the same plot when I enter different names in trajectory Display GUI using a + symbol and what is the minimum number of tdump files which I need to create a cluster?
  • Click for answer.How do I estimate the absolute (in km) and relative (%) errors when using the HYSPLIT trajectory model? Where does the error mostly come from (e.g., truncation error, interpolation error or wind field error)?
  • Click for answer.What are the levels of uncertainty associated with back trajectory calculations in HYSPLIT?
  • Click for answer.What are the levels of uncertainty associated with back trajectory calculations of ambient and potential temperature in the Hysplit model meteorological outputs? I am trying to use the model, to do 5-day backward trajectories of two air parcels (200m and 1000m) starting at my study location. I would like to use the temperature outputs to understand processes within my study region and would like to know its accuracy. (April 2004)
  • Click for answer.How does HYSPLIT consider the vertical coordinate. If a trajectory meets the surface how does HYSPLIT work? Does it reflect, or use another method to solve it?
  • Click for answer.We calculated three isentropic back trajectories simultaineously for the same latitude and longitude, but for three different altitudes. Then we calculated one back trajectory by itself. To my surprise the results are different. Why are the back trajectories different if you calculate them jointly or separately?
  • Click for answer.Is the effect of the boundary layer considered when the model computes the trajectory?
  • Click 
	              for answer. I have been running trajectories every hour for 24 hours to get a general idea of the airmass origin a single point during this period. Is there a way to obtain an average visual trajectory for a 24 hour period?
  • Click for answer. Why are there small differences in my trajectories when I run more than one at a time versus running just one trajectory?

Interpretation of Concentration Results

  • Click for answer.Is it possible to interpolate between two lines of equal concentration or between a maximum and a line?
  • Click for answer.How is concentration defined in HYSPLIT and how does it relate to the radiation units of Bacquerel?
  • Click for answer.Is there any way to obtain the concentration and deposition information in a data file instead of a graphic?
  • Click for answer.When running HYSPLIT concentrations I got 10E-11 ... 10E-17 for values of concentration. I understood that the unit is mass/ My question is that in order to apply the numbers to total emission from a power plant stack, how can I calculate real plume concentration? Suppose it is 167381 ton/yr and converted 19 ton/hr, then, can I just multiply 19 ton/hr by 10E-11... to get real level?
  • Click 
              for answer. I found a case where there are deposition contours where there are no corresponding air concentration contours. Why?

Return to top

Running HYSPLIT on the READY Server


  • Click for answer.How do I calculate and visualize backward and forward trajectories for our air pollution monitoring station on the READY website?
  • Click for answer.My research is to use wind trajectories to find pollutant sources in Chicago. How do I translate longitude and latitude into the exact location, say Sears Tower or United Center?
  • Click for answer.What is the longest period over which trajectories can be calculated on the web?
  • Click for answer.Is it possible to extract the the trajectory-data into an ascii-file?
  • Click 
	              for answer.I am trying to do a backward trajectory for parcels involved in historical weather events, and I know that even minor changes in the point or time can yield much different results for the parcel origin. Is there a way to do an "ensemble" run, slightly perturbing the starting point?


  • Click for answer.The model always gives a deposition plot. Can this influence the concentration plot? I mean: If you have a gas with no deposition at all, is the the concentration plot then still valid?
  • Click for answer.Does HYSPLIT output 3 dimensional concentrations?


  • Click for answer.How come when I use the back button on the browser, my settings to the HYSPLIT Model Setup form are returned to the default settings?
  • Click for answer. Which datum do you use for the output from your models? NGVD 29? NAD27? NAD83? WGS84? Clarke Spheroid? And what does datum mean, anyway?

Return to top

Running HYSPLIT on your PC

  • Downloading, Installing and Running the Model

  • Click for answer.Where can I download the HYSPLIT PC version?
  • Click for answer.What RAM and HD requirements are needed for running the trajectory and dispersion models on the PC?
  • Click for answer.Is it necessary for the CONTROL file to be named "CONTROL"?
  • Click for answer.How do you specify more than one meteorological file in the 'set Meteo_file' line of the Auto_traj.tcl script so the loop can run through many months of data?
  • Click for answer.Can I produce the particle display on the PC version?
  • Click for answer.Why do I get this message when trying to run PC Hysplit: "C:\hysplit4\guicode\hysplit4.tcl is not a valid Win32 application"?


  • Click for answer.How do I run a BACKWARD trajectory using the PC version of HYSPLIT?
  • Click for answer.When I ran a backward trajectory from a location on a previously computed forward trajectory, I did not return to the starting location of my forward trajectory. Why?
  • Click for answer.Is there a way to run the model in "batch mode" that would generate trajectory endpoints?
  • Click for answer.I am running a statistical program that groups hundreds of trajectories over one another and creates probability fields. My problem is that generating one trajectory at a time and converting the ASCII trajectory endpoint data into Excel format is far too time consuming. Is there a way to run a number of trajectories at once?
  • Click for answer.An FAQ question under HYSPLIT in General explains how to adjust the start height to account for differences between model terrain and actual terrain. I'm assuming that it's most accurate to use actual elevation plus my desired trajectory height as a starting height, and specify the option KMSL=1 in the setup.cfg file on the PC. Is this correct? Occasionally these trajectories are wildly different f rom those with KMSL=0 and the start height specified as meters above ground level.
  • Click for answer.Is it possible to represent in the same map both forward and backward trajectories in order to make a comparision of the integration time step?
  • Click for answer.For the web version of HYSPLIT, we can get temperature, potential temperature, precipitation, etc., along the trajectory. Can we do the same with the PC version?
  • Click for answer.I purchased NCEP reanalysis GRIB data for summer 2001, downloaded the PC HYSPLIT model, and figured out how to convert the GRIB format to the HYSPLIT format using the compiled fortran program: ncr2arl. However, I am unsure about what latitude and longitude to input. I read that lat and lon are at the center of a 100x100 grid that is extracted from the GRIB file. So if my study area is around the North Eastern Pacific/West Coast of the US, should I just input my center lat and lon from my study area? I am trying to do backward air parcel trajectories of the marine atmospheric air located off Oregon. I am worried that if the grid extracted is only 100x100 from my center lat/lon that long duration backward trajectories might cover a larger area than the grid extracted. How should I deal with this? Finally, when using display.exe (or HYSPLIT GUI interface) to display contour maps, should I again enter the center lat/lon of my study area (in other words is the diplay.exe plotting the extracted 100x100 grid?) and what is meant by the radius (degLat)? What value of radius should I put in and how does different values change my contour map?
  • Click for answer.When I set up a matrix run and have the heights specified at a certain altitude (using MSL), why do the results show the starting points at different heights instead of at a constant one? (April 2004)
  • Click for answer.Please let me know is it possible to classify the trajectories by cluster analysis and draw a mean trajectory of each cluster using the PC version of Hysplit.


  • Click for answer.The model prints on a large scale (all of Europe), while the concentration area itself is sometimes rather small. Is it perhaps possible to reduce the scale of the map?
  • Click for answer.I have several months of the EDAS data files on my PC and wish to run several months at one time. Is there a way to run the model in "batch mode" that would generate deposition and concentration for each endpoint?
  • Click for answer.I downloaded the latest version of HYSPLIT for the PC. How do I run the dispersion simulation backwards? I tried a negative number of hours run time, as for backward trajectories, but I got the same result as with positive number of hours.
  • Click for answer.I am trying to run HYSPLIT (PC version) for Mercury (emit in gas phase). For chemical depositon set up, the model asks f or Surface Reactivity Ratio, Diffusivity Ratio, and Effective Henry's Constant. At this time I could not find any of the se properties for (Hg+2 or HgO). I wonder is there anyone who did this pollutant simulation before. Any suggestion about where to obtain this information would be appreciated.
  • Click for answer.I am looking to run concentrations for about 300 plants in HYSPLIT for use in an econometric analysis. I am running the concentration model for each plant and producing a GIS map for each model output. My problem is that generating one concentration and one GIS map at a time is very time-consuming and I am wondering if there is a way to do all 300+ of the concentrations at the same time? I want the model (time, pollutant characteristics, etc.) to be all the same for each model run, the only difference is the emission source and the height.
  • Click for answer.How do I calculate deposition?


  • Click for answer.I am using the PC version of the model on a Windows NT computer and am having trouble running it. I downloaded some archived files from the ftp site, and when I ask the model to use them I get the following error:

    HYSPLIT4 (Dec 98) - Initialization
    Unable to find file: JUN98.BIN
    On local directory : C:/HYSPLIT4/METDATA
    Check input CONTROL file for correct values
    Stop - Program terminated.

    I edited the control file by taking out the relative reference pathfile and put in the specific fully qualified directory path C:/hysplit4/metdata/ Do you have any suggestions on how to correct this problem?
  • Click for answer.I purchased FNL meteorological files from NCDC for 1998, and they sent me a tape with the files. When I run the HYSPLIT model however, I get the following message:
  • Click for answer.I have been quite successful running the trajectories off the website, but have not been able to get the downloaded software to work correctly. It seems to put out the default example graph every time, even after other meteorological data has been entered. Since I have been using HYSPLIT on the web, it doesn't really matter. But I was wondering if the website accesses the hard drive, or if the software is actually not necessary and the website can be used instead? Are there things I can do using the software that are not available on the web? If so, what?
  • Click for answer.Why can't I run a back trajectory from 1900 UTC on 31 July, 2001 using downloaded FNL data? I attached my CONTROL file for you.
  • Click for answer.When using trajplot.exe to plot multiple HYSPLIT output files, we are having trouble getting trajplot to recognize output files unless they are named tdump, tdump1, tdump2, ..., tdumpn. Is it necessary for the file name to contain "tdump"?
  • Click for answer.I am trying to calculate trajectories using my own MM5 output files. I compiled the mm5toarl routine and created the arl formatted files compatible with HYSPLIT, but it still doesn't work. I used the chk_file program and there is an error with the header. Should I change something in the mm5toarl routine?
  • Click for answer.I am attempting to run the HYSPLIT model daily for a 10 year period. I received the error message:

    	"*ERROR* runset: Numb meteo grids exceed DEFGRID limit -              12"

    Is there a restriction of the length the model can run for? If so how do I get around this or do I have to restrict it to a single year?
  • Click 
              for answer.I have been trying to use the convert program to trim the white space around the edge of the hysplit plots using the ImageMagik 'convert -trim' command. This seems to have no effect on the image. The '-crop IxJ' option has limited success. Is this an underlying issue with the .ps files that trajplot and concplot create?
  • Click for answer.I received the following error. What could be wrong?
    	> Model started ...
    	> HYSPLIT48 (March 2007) - Initialization
    	> Calculation Started ... please be patient
    	> *ERROR* metpos: start point not within (x,y,t) any data file
    	>         xp,yp :    7.800000       12.60000    
    	>    time (min) :     53864640
  • Click for answer. Why are the results different for the same run on the web versus the PC?
  • Click for answer. When I "Run Standard Model" after saving the "Trajectory Setup", I get a static message
    		> Model started ...
     	> HYSPLIT48 (February 2008) - Initialization

    This does not change. Why?

Return to top

Volcanic Ash Model (currently HYSPLIT, formerly VAFTAD)

  • Click for answer.As an airline pilot I always check the volcanic ash dispersion forecasts for Colima and Popoctptl whenever I fly Toronto-Mexico City, however I'm a little uncertain as to just how much gravity I should treat them with. If I was flying, say, from Miami to Mexico City at 18000ft can you tell me just what I might encounter if VAFTAD has a big red "blotch" along my route? Should I be planning to avoid it altogether or is the ash likely to be too dissipated to cause a problem?
  • Click for answer. The HYSPLIT distribution models will often show significant areal coverage. Is there a threshold particle density that defines the plume boundaries and if so, is it one that corresponds to a known aircraft danger/non-danger threshold ?
  • Click for answer. I am doing a project on air pollution due to Volcanic activity. I am missing one major part of the project: a model for the plume so that I can calculate the contaminant dispersion. I was wondering if one of you were familiar with such models and if I could get any information about them.
  • Click for answer. Is it true that the VAAC's do not issue volcanic ash dispersion model forecasts (sometimes called VAFTADs using the name of an earlier model), but use them as guidance, like the PUFF and AVHRR? If I wanted to see your forecast, then would need to look up your bulletin and the attached VAAC chart along with the SIGMETS bulletin, or do you send the VAFTAD out as a graphic?
  • Click for answer. I saw a program on TV that stated volcanic gas and/or ash is necessary for earth's healthy atmosphere, and that volcanic erruptions cause a protective layer between us and the sun. Is this in fact true?

Return to top

Modified: September 16, 2011
Dept. of Commerce | NOAA | NOAA Research | Website owner: ARL