check areal weight calculation

[gnodnooh]

Hi @Shrad-Shukla1 and @Yoon-Eric

(edited on April 30)

Okay, (xr_da * aw_factor).mean(dim=[lon_name, lat_name]) calculates the areal-weighted average per grid cell. see below example.

libraries/python_functions/spatial_aggregation.py

Lines 297 to 324 in 912606b

	def calc_spatial_mean(
	xr_da, lon_name="longitude", lat_name="latitude", radius=EARTH_RADIUS
	):
	"""
	Calculate spatial mean of xarray.DataArray with grid cell weighting.
	Parameters
	----------
	xr_da: xarray.DataArray
	Data to average
	lon_name: str, optional
	Name of x-coordinate
	lat_name: str, optional
	Name of y-coordinate
	radius: float
	Radius of the planet [metres], currently assumed spherical (not important anyway)
	Returns
	-------
	Spatially averaged xarray.DataArray.
	"""
	lon = xr_da[lon_name].values
	lat = xr_da[lat_name].values
	area_weights = grid_cell_areas(lon, lat, radius=radius)
	aw_factor = area_weights / area_weights.max()
	return (xr_da * aw_factor).mean(dim=[lon_name, lat_name])

[Shrad-Shukla1]

Hi Donghoon, Yes, let's test those functions. However when I had compared my results using the same function with what Chris Funk had generated, they both resulted in the same mean values over different parts of Africa. Thanks, Best regards, Shrad Shraddhanand Shukla Researcher, Climate Hazards Center, UCSB. <https://chc.ucsb.edu/people/shraddhanand-shukla> Office Phone: (805) 893-3261 Co-author: Drought Early Warning and Forecasting <https://www.sciencedirect.com/book/9780128140116/drought-early-warning-and-forecasting> ORCID profile <https://orcid.org/0000-0003-0077-6733>; Google Scholar <https://scholar.google.com/citations?user=wtJoeUIAAAAJ&hl=en>; Scopus <https://www.scopus.com/authid/detail.uri?authorId=23978908300> *Recent publications:* 1. Contrasting Performance of Panel and Time-Series Models for Subnational Crop Forecasting in Sub-Saharan Africa <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4635817> (preprint) 2. Using Earth Observation Products to Predict Maize Prices in Southern Africa <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4369127> (preprint) 3. Investigating the Strength and Variability of El Niño–Southern Oscillation Teleconnections to Hydroclimate and Maize Yields in Southern and East Africa <https://journals.ametsoc.org/view/journals/hydr/25/1/JHM-D-23-0098.1.xml> 4. Anthropogenic Climate Change Negatively Impacts Vegetation and Forage Conditions in the Greater Four Corners Region <https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022EF002943> 5. Investigating the relationship between growing season quality and childbearing goals <https://www.sciencedirect.com/science/article/pii/S0959378023000432?via%3Dihub>

…

On Mon, Apr 29, 2024 at 6:49 PM Donghoon Lee ***@***.***> wrote: Hi @Shrad-Shukla1 <https://github.com/Shrad-Shukla1> and @Yoon-Eric <https://github.com/Yoon-Eric> I believe the calc_spatial_mean function works here when the data includes the equatorial region (latitude = 0). Once you "crop" the data to the non-equatorial regions, the aw_factor yields weights that are less than 1.0 for most grid cells. This results in consistently lower areal-weighted averages than the actual values. Happy to discuss more. https://github.com/chc-ucsb/libraries/blob/912606b66af96833d3d61d8b77c29b3a3667b813/python_functions/spatial_aggregation.py#L144-L162 https://github.com/chc-ucsb/libraries/blob/912606b66af96833d3d61d8b77c29b3a3667b813/python_functions/spatial_aggregation.py#L297-L324 — Reply to this email directly, view it on GitHub <#2>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKD6S3QMM7VO3T2BVYOWIF3Y732DDAVCNFSM6AAAAABG7LIG6WVHI2DSMVQWIX3LMV43ASLTON2WKOZSGI3TAMZSHE4TGNQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[gnodnooh]

Here is an example. What we are calculating is the average precip over 9 grid cells.

[Shrad-Shukla1]

Ok I see. So I guess what we need to do is simply divide the weights by the maximum weight in a given polygon. As of now the weights are global. Does that make sense? Thanks for catching this. Best regards, Shrad Shraddhanand Shukla Researcher, Climate Hazards Center, UCSB. <https://chc.ucsb.edu/people/shraddhanand-shukla> Office Phone: (805) 893-3261 Co-author: Drought Early Warning and Forecasting <https://www.sciencedirect.com/book/9780128140116/drought-early-warning-and-forecasting> ORCID profile <https://orcid.org/0000-0003-0077-6733>; Google Scholar <https://scholar.google.com/citations?user=wtJoeUIAAAAJ&hl=en>; Scopus <https://www.scopus.com/authid/detail.uri?authorId=23978908300> *Recent publications:* 1. Contrasting Performance of Panel and Time-Series Models for Subnational Crop Forecasting in Sub-Saharan Africa <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4635817> (preprint) 2. Using Earth Observation Products to Predict Maize Prices in Southern Africa <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4369127> (preprint) 3. Investigating the Strength and Variability of El Niño–Southern Oscillation Teleconnections to Hydroclimate and Maize Yields in Southern and East Africa <https://journals.ametsoc.org/view/journals/hydr/25/1/JHM-D-23-0098.1.xml> 4. Anthropogenic Climate Change Negatively Impacts Vegetation and Forage Conditions in the Greater Four Corners Region <https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2022EF002943> 5. Investigating the relationship between growing season quality and childbearing goals <https://www.sciencedirect.com/science/article/pii/S0959378023000432?via%3Dihub>

…

On Mon, Apr 29, 2024 at 7:23 PM Donghoon Lee ***@***.***> wrote: Here is an example. What we are calculating is the average precip over 9 grid cells. image.png (view on web) <https://github.com/chc-ucsb/libraries/assets/12818281/fa4f1332-d300-419b-bfca-755c9c1b643c> — Reply to this email directly, view it on GitHub <#2 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKD6S3UP5XWIU2AE74XHYPDY736CNAVCNFSM6AAAAABG7LIG6WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOBUGIYDSMJUGU> . You are receiving this because you were mentioned.Message ID: ***@***.***>