Calculating pixel-wise mean from ImageCollection

Question

Suppose I run the following code to get an ImageCollection in Google Earth Engine:

collection = ee.ImageCollection("ECMWF/ERA5_LAND/HOURLY").select("total_precipitation")
collection = collection.filterDate("2020-01-01", "2020-01-02")
print(collection.size().getInfo())

This outputs 24, because the data is hourly and it is a collection of images for 1 day. If I then run the following code to inspect the first image:

geometry = ee.Geometry.Polygon([
    (-119.6863362607843, 37.91032228872993),
    (-119.24681336572189, 37.91032228872993),
    (-119.24681336572189, 37.58602618425783),
    (-119.6863362607843, 37.58602618425783),
    (-119.6863362607843, 37.91032228872993)
])

img = collection.first()
count = img.reduceRegion(ee.Reducer.count(), geometry=geometry, scale=1000)
count.getInfo()

This outputs:

{'total_precipitation': 540}

Which I believe means there are 540 pixel values in that band's image for that region.

How would I go about getting a single composite image for the whole day's data, which is an average of the 24 hourly values? To clarify: from the ImageCollection above I want to produce a single Image with 540 pixel values, where each value is the mean of the 24 hourly values. Is this possible?

Answer 1

It's unclear what you mean by "the underlying hourly images each have the same dimensions (in this case 4x6 pixels)". Each total_precipitation band in the ECMWF/ERA5_LAND/HOURLY collection contains a single value per pixel. There are 24 images per day, but they don't have a 4x6 structure; just one value per hour. When you take the mean of those, you end up with a single value for the whole day.