By Andy May
It came up in the comments on my last post, CERES Albedo. What is the best way to compute Earth’s albedo? The CERES data is supplied as a 1° x 1° latitude/longitude grid. It is widely accepted that Earth’s global mean albedo is around 30%. The question is then: What is the best way to estimate it using the CERES satellite data? There are two basic ways. One is to use the average solar radiation arriving at the top of the atmosphere (CERES EBAF variable “solar_mon”), which is about 340.2 W/m2 and divide that into the total solar shortwave radiation (SW) leaving (reflected from) the Earth (toa_sw_all). Using these two numbers we get an albedo of about 29%.
The second way is to compute the albedo for each of the 64,800 one-degree latitude & longitude cells and then compute the area-weighted global mean of all the albedo calculations. When this is done, the albedo is 31.3%. Statistically this is the same as the 29% calculation because the errors in measuring solar_mon and toa_sw_all are large (> ±2 W/m2), plus we do not know how much solar longwave radiation (LW) is reflected, but the problem is worth examining. Figure 1 shows the elements. Click on it to enlarge it and show it in full resolution.
The spreadsheet on the left of figure 1 shows the area-weighted yearly means for the CERES outgoing SW and the incoming solar radiation. Dividing the first column by the second results in the last column, labeled “conventional gm albedo.” The basic calendar year cell-by-cell area-weighted albedo global average albedo is next and labeled “cbc albedo.” The next column (“cbc rm36 albedo”) is computed by taking a 36-month running mean (centered) of both toa_sw_all and solar_mon, then computing a month-by-month and cell-by-cell albedo, then extracting an area-weighted global mean albedo from that dataset for each year. In terms of yearly global mean albedo, it matches the year-by-year and cell-by-cell calculation closely.
The set of maps in the middle of figure 1 show that the two cell-by-cell albedo calculations are very similar for 2025. The simple “SW out/solar in” calculation is the same value for every cell and the important detail we see cell-by-cell is hidden in the global mean.
The right-hand maps and graph show the 25-year trends that result from the two ways of computing the cell-by-cell albedo means. The upper trend map shows areas of decreasing albedo in either light yellow or blue. Areas of increasing albedo are shown in orange to red. The year-by-year albedo changes in the upper trend map are plotted in red on the graph at the bottom right of figure 1.
The middle right trend map is the trend in albedo after taking a 36-month centered running average. Notice it is almost a mirror image of the upper year-by-year trend map. Taking the 36-month running mean has offset the very active albedo data and reversed its slope, as shown by the blue line in the lower right corner of figure 1.
Conclusions
Essential details of the global albedo distribution are lost when using global averages as is done in the conventional calculation. Taking a running average of either the components of the albedo calculation or the computed albedo causes a shift and a change in slope in the albedo trend.
The best way to compute global albedo is to do it cell-by-cell and then make an area-weighted global mean of the cell-by-cell albedo values. I prefer to use calendar yearly means to remove seasonality because running means distort the trends. This means the best estimate of albedo, using CERES data, is 31.3%. This is also the best way to determine the albedo trend (the red line in the graph).
The Sun’s position in the sky changes constantly, so a snapshot momentary conventional global albedo is useless. At least the “CERES albedo” has a constant frame of reference, it may not be the exact albedo as measured from space, but it can be compared from month to month because the incoming and outgoing radiation are (at least theoretically) always from the same reference angle. The only changes (ignoring orbital drift and other sources of instrument error) are the albedo components on the surface (clouds, ice, etc.). At least the “CERES Sun” is not constantly moving.
Not sure if the graphics will show up in this. If not, the original comment is at
https://wattsupwiththat.com/2026/05/22/ceres-albedo/#comment-4197978
Thanks for all your great work,
w.
===
Thanks, Andy. In general, taking averages of gridcell ratios as you are doing is less accurate than taking the average of the numerator and the denominator and then taking the ratio of the two.
I asked perplexity.ai to explain this. See below.
Best to you,
w.
===
Here’s my question:
I have a gridded global dataset of average TOA reflected sunshine, and another of gridded sunshine.
Albedo is the ratio of the two, reflected/sunshine
Which is more accurate—
a) take the weighted mean of the reflected and the weighted mean of the sunshine, and divide the first by the second, or
b) divide the gridded reflection by the gridded sunshine, and take the weighted mean of the albedo of the individual gridcells
The first method gives ~ 0.29, and the second gives ~ 0.31
Here’s the answer:
===
Method (a) is the mathematically correct way to get the global mean albedo: compute the area weighted mean reflected flux and the area weighted mean incoming flux, then take their ratio.
Method (b) generally gives a biased result unless every gridcell has the same incoming sunshine.
Why (a) is the right global mean
Define for each gridcell (i):
· (R_i): time mean reflected shortwave at TOA (e.g. W/m²)
· (S_i): time mean incoming shortwave at TOA (same units)
· (A_i): area of the gridcell
Then by definition, the global mean planetary albedo is
https://wattsupwiththat.com/wp-content/uploads/2026/05/albedo-1.png
This is exactly your option (a): area weighted mean (R) divided by area weighted mean (S).
Why (b) differs
Option (b) computes per cell albedo (α_i = R_i / S_i), then takes an area weighted mean of (i):
https://wattsupwiththat.com/wp-content/uploads/2026/05/albedo2.png
This equals α_global only if (S_i) is constant across all gridcells (or uncorrelated with (α_i) in a very particular way), which is not true for Earth because insolation varies strongly with latitude, season, and clouds. High latitude and cloudy regions tend to have different (S_i) and (α _i) than low latitude clear regions, so a simple area average of (α_i) overweights regions with low (S_i) relative to their contribution to the global energy budget.
A simple toy example:
· Cell 1 (tropics): (S_1 = 400), (α_1 = 0.25) → (R_1 = 100).
· Cell 2 (high lat): (S_2 = 200), (α_2 = 0.50) → (R_2 = 100).
· Equal areas.
Then
https://wattsupwiththat.com/wp-content/uploads/2026/05/albedo-3.png
but the simple area mean albedo is
https://wattsupwiththat.com/wp-content/uploads/2026/05/albedo-4.png
Option (b) overestimates the global mean because it gives the low insolation, high albedo region the same weight as the high insolation, low albedo region, even though it reflects less absolute power.
Interpreting your 0.29 vs 0.31
Your ~0.29 from method (a) is consistent with commonly cited estimates of Earth’s planetary albedo (~0.29–0.31, depending on period and dataset). The ~0.31 from method (b) is exactly the kind of upward bias you expect when you area average per cell albedos without weighting by incoming flux.
So:
· Use (a) for global mean planetary albedo relevant to energy balance.
· Use (b) only if you explicitly want the unweighted area mean of local albedos as a descriptive map statistic, not as an energy correct global albedo.
===
I still think we are probably talking past each other and this is more a matter of definitions than anything else. You seem to be emphasizing “global albedo” in the sense of Earth as a point with an observer out in space measuring the light hitting the point and the light leaving and taking the ratio. I agree that case is probably closer to 0.29.
My point is that generally on Earth, we are more interested in the global mean albedo and how it changes with time. That is the distribution of albedo because that is what affects the climate. The angle the sun makes relative to Earth’s surface changes constantly and has a huge effect on climate also. Thus, the changing distribution of albedo should, and probably does, as in glacial periods.
Here is a better summary of what I’m thinking, I put it on WUWT also:
To summarize the albedo debate from my perspective.
The difference between our albedo estimates comes down to definition and purpose.
Your “point‑source” albedo is correct for the specific instant and geometry at which the Earth–Sun system is viewed from far away. But that value is not climatically meaningful because the solar zenith angle changes continuously—minute‑to‑minute, season‑to‑season, and over orbital cycles.
To study climate, what matters is global mean albedo, computed over the full CERES grid, weighted by area and integrated over time. This quantity tells us how much of the incoming solar energy the Earth system reflects on average, and how that reflection changes spatially and temporally.
A single‑instant whole‑Earth albedo seen from deep space might be ~29%, depending on the viewing geometry and time of year. But the global mean albedo, averaged properly over the full illuminated hemisphere and over time, is closer to 31% in CERES EBAF.
So, the disagreement isn’t about arithmetic—it’s about what albedo means in a climate context. Point‑source albedo describes a momentary optical property. Global mean albedo better describes the Earth’s energy balance over long periods of time. It is the albedo measure that is important climatically.
Hi Andy, the variation in albedo over the areas of ocean that experence high surface pressure in the SH in summer is of interest because areas of low albedo will likely expand and contract according to the extent of uplift within the high latitude zones of low surface pressure in the opposite winter hemisphere, the Aleutians in particular. Insolation is 6% stronger in January and penetrates to depth.