0

i do have soil temperature data (5 cm depth) from 1991 and 2021 for Germany, and wanted to interpolate the data for each year as a class of 3 seasonal months (winter = 12, 1, 2; summer = 6, 7, 8). Unfortunately, the spatial distribution of the data for 1991 is bad, lacking several data points for east Germany.

Data Summer 1991

How would i proceed using the geostatistical wizard in ArcGIS Pro?

I tried to use Simple Kriging, but the error was too big and overall, the result was lackluster.

However, the method of using simple kriging should be fine, considering the mean, median, skewness and kurtosis matched the criteria, or am i thinking wrong?

Is the data even usable for interpolation?

Improve this question
2
  • Usable? That depends on what you're trying to prove. Accuracy is a etherial concept, it would be useless if you were expecting city sized accuracy but might work for a country sized product and then disclaimer the heck out of it.
    – Michael Stimson
    Commented Jul 13, 2023 at 6:55
  • I see, but of course i would like to get the best results i possibly can, with the tools i have. Would a dataset like this rather use ordinary or simple kriging? And is there a way to visualise the error of areas (prominently east germany) within ArcGis? The main purpose of this work is for my studies, where we have to find suitable data and interpolate them. Ideally i would want to discuss the differences of soil temperature between 1991 and 2021, hence the statistically important time difference of 30 years.
    – vaber32
    Commented Jul 13, 2023 at 7:05

0

Reset to default

Your Answer

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