I have downloaded daily precipitation data from Global Weather data website and want to create an annual precipitation map like this:
example of desired map
(source: worldatlas.com)

The downloaded data is in TXT and CSV format. The area I am covering has 96 weather stations, so I have 96 files. I am totally confused how to assemble and use this data for the said map.

I am using Arcgis 10.

  • 1
    How do you want to display it? Like points at the weather stations' locations with some symbology relative to precipitation there, or like surfaces interpolated over your whole area? No matter what, you need to do some data handling first. Do you want daily mean precipitation, monthly mean, the value each day...? Please update your question with more information on what your goal is. Also include what software you have access to.
    – Martin
    Commented Aug 29, 2014 at 10:26
  • I have added much more information in my question.
    – Mehnaz
    Commented Aug 29, 2014 at 10:57

2 Answers 2


What you need to do to create a continuous surface representing precipitation is a process called interpolation. ArcMap has a number of tools to do this, based on a variety of statistical and sampling approaches. I'd recommend inverse distance weighting (IDW) as a starting point, because it's one of the simplest to use.

The input for IDW is a single feature layer of points (in this case, weather stations). Each point needs to have one value (i.e., mean annual precipitation), and then the interpolation function estimates what the values should be in between the known points.

Importing station data into ArcMap is straightforward as long as you have a location (latitude/longitude) for each one. The tricky part here is that each station currently has many (see note below) values, and you want one map. You can do an interesting but complicated approach after importing the data, or you can summarize the data using another piece of software (I'd use Excel, but many statistical packages would give the same results) before putting it into ArcMap.

Note: One minor consideration is how much time the data covers. Is it daily precipitation for one year? If so, your resulting map will be "precipitation for 2013", not average annual precipitation. If it is 1961-1990 (or some other interval), then you need to group and summarize the precipitation by year.

  • 1
    The dedicated toolbox for interpolation is Geostatistical Analyst. You also have interpolation tools in the 3D Analyst and Spatial Analyst toolboxes.
    – GISGe
    Commented Aug 29, 2014 at 12:45
  • Ah, ok. I've only ever done it through Spatial Analyst :)
    – Erica
    Commented Aug 29, 2014 at 12:55
  • It's more than fine if you're not a geostatistician, but you don't absolutely need Spatial Analyst, the exact same tools are available with 3D. And much more interpolation tools and options are available within GeoStatistical Analyst
    – GISGe
    Commented Aug 29, 2014 at 12:59

Regarding the interpolation tool from ArcGis I must say that I got different outputs using the available methods (IDW, Kriging, CoKriging, Kernel Smoothing).

In order to get a reliable output there are many physical parameters to consider:

  • Gauge areal distribution
  • Topography
  • etc.

I'm sure there are some more, these are on top of my head but don't forget that each method has many other parameters that can largely modify the output (e.g. changing the bandwidth from the Kernel Smoothing method).

I think that's all I wanted to say. I just wanted to be informative with my answer and share some experience.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.