I'm working with Google Earth Engine images from ERA5-Land Hourly and am trying to take the mean or the sum of the hourly data on a daily basis. I've come across an issue where taking the mean seems to multiply the resolution by 10 and I'm at a loss trying to understand it.
Below is some Python code for a minimum reproducible example of the issue. Can anyone tell me what I'm doing wrong here?
import ee
# Select the ImageCollection:
polygon = ee.Geometry.Polygon([
(-119.6863362607843, 37.91032228872993),
(-119.24681336572189, 37.91032228872993),
(-119.24681336572189, 37.58602618425783),
(-119.6863362607843, 37.58602618425783),
(-119.6863362607843, 37.91032228872993)
])
ic = ee.ImageCollection("ECMWF/ERA5_LAND/HOURLY").filterDate("2020-01-01", "2020-01-10").filterBounds(polygon).select("total_precipitation")
# Take a sample image:
num_img = ic.size().getInfo()
img_list = ic.toList(num_img)
img = ee.Image(img_list.get(0))
# Print the scale:
print(img.projection().nominalScale().getInfo())
This outputs 11131.949079327358 - which is the expected resolution for ERA-5 Land.
Now if I take the mean of the first day's images, and print the scale, it has multiplied by 10:
img = ic.filterDate("2020-01-01", "2020-01-02").mean()
print(img.projection().nominalScale().getInfo())
Outputs: 111319.49079327357
Note that this happens if I take the first day, 2 days, etc. What am I doing wrong?