We have ERA5 time series data which was used to extract maximum and minimum temperature. As the spatial resolution was not good therefore we downscaled it to 90 meters to increase the correlation coefficent value against in-situ data.
so i wanted to know whether this step is scientifically correct or not???
Hi Gomal,
If I understood you correctly, you up-sampled the dimensions of the temperature data from a single pixel dimensions of 0.25x0.25 degrees to ~90x90 meters.
IMO there is nothing wrong with doing that, even though no new information is added. Depending on the interpolation method used in the up-sampling, you will achieve a "smoother" correlation function, that will potentially allow finer determination of the correlation peaks . The downside is saving a lot of redundant information that cam potentially affect performance and disk usage.
Good luck :)
We have downgraded to 900 m and information resulted has been very useful to reinforce our databases and geographical information systems
You would need to make an elevation correction for any inhomogeneous pixel. This is described in section 5.2 of Book Best Practices Handbook for the Collection and Use of Solar ...
In order to be sure, you can receive and compare the data related to the meteorological stations in the desired time period.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing