In literature, eight subregions of coherent seasonal rainfall anomalies have been identified for a certain region. I would like to construct time series for these eight regions from a netcdf file with precipitation data. I do not have the exact latitudinal and longitudinal bounds but I do know the regions. Do I ought to select a region (square or rectangle) which is representative of the region and slice the data? My main issue here is that the subregions are irregularly shaped.
Hello, do you have lat/lon values in your netcdf file? If yes then the task is simple: you need to spend some time defining the masks for your zones in a regular grid (e.g. 100x200 points with 0.1 deg resolution) - in this approach it doesn't matter whether your regions are regularly shaped or not. They can be of any shape, have holes, etc. You can create the masks manually or using some script. You just have to define some key points and rules to fill the area. For example, here are two steps for Greenland:
http://puu.sh/AZZLE/33d46ae5aa.png
https://puu.sh/ARVg3/736b1b1357.png
Then you have to create a table of the same size (100x200) and fill it with precipitation values from netcdf. After that, the product of this table with the corresponding mask will give the result you're looking for.
Three main techniques are conventionally employed to average spatial rainfall; (a) the common arithmetic average; (b) the Thiessen polygon; and (c) the Isohyet method. A works well for regions with no changes in topography. B works well for a region with eny subregions similar to the full region. C works better for changes in altitude abive sea level. As you know all subregions, classify them in a, b, or c and carry out the analysis accordingly. The Math to carry the analysis is Pm= (Pa+Pb+Pc+Pd+Pe+..Pn)/n for a; where Pm= mean precipitation, Pa, Pb..= Precipitation of subregions a, b,..For b, use: Pm=Sum((Pa*Aa)*(Pb*Ab)+(Pc*ac)+(Pd*Ad)*..+(Pn*An))/A; where A = area of the full region; Aa = Area of the subregion a. For the Isohyet method you have to discretize Areas according the altitud above sea level and calculate mean precipitation for each area say between 1900 and 2000 masl. Then use the equation descibed immediately above.
Hope these statements help you with your evaluations.
Kind Regards
Sincerely
Jose Navar, Ph.D
If u have arc GIS you can use interpolation other software as well can interpolate rainfall in the catchment. Once I try to produce rainfall in basin using stations rainfall using Thessisen polygon its shows error of extrapolation in catchment. Other methods for interpolating rainfall are Isohyet. I support @Jose Navar answer above.
Thank you all. I ended up using QGIS to create shapefiles. I then used some code to mask the data
- Badges
- Science topic
- Similar topics
- Engineering
- Computer Engineering
- Computer