Separate concentration from emissions

A concentration describes the composition of air being observed. A surface flux describes transfer into or out of the atmosphere over a defined area and time. Wind can carry gas away from its source, and natural exchanges also contribute to the observed distribution. A high concentration above a location is therefore not a direct meter reading of that location’s emissions.

Consider an original thought experiment: two identical sources release the same amount of gas, but one plume is rapidly dispersed by wind while the other remains concentrated nearby. The observed concentration patterns can differ even though the release rates are equal. Inferring emissions requires a model of how the gas was transported, not just a ranking of the brightest map pixels.

NIES describes higher-level GOSAT flux estimates as inverse calculations combining concentration observations, other measurements, meteorological information and an atmospheric transport model. The analysis works backward from observed atmospheric conditions to possible source and sink strengths. Its assumptions and uncertainty are part of the result.

In an original bookkeeping example, a region releases 10 units of carbon per day and removes 7 through other processes. Its net contribution is 3 units per day under those definitions. Reporting only the net value would conceal the larger opposing flows. Likewise, a modeled net flux should not automatically be labeled as a complete inventory of fossil-fuel emissions.

A concentration map is not an emissions meter

A NIES dataset released on April 14, 2026 uses GOSAT observations in the NISMON-CO2 framework. Its documented calculation holds the fossil-fuel emissions input fixed while optimizing other flux components. The output’s one-degree grid also differs from the coarser transport-model grid. A fine output grid does not mean every cell was independently measured by the satellite.

For an original review exercise, someone claims that this dataset independently corrected every factory’s fossil-fuel inventory. The documented fixed input contradicts that interpretation. A better summary states which observations constrain the model, which quantities are adjusted, which are held fixed and which spatial scales the analysis can support.