China Gas Apparent Sector Demand Model

What this model actually allocates

The headline bars below show modeled apparent demand split into the four sectors for the selected year. The total of every bar equals JODI apparent demand for that month — by construction, the four sectors sum back to the published monthly total.

That means each sector value is not a meter reading. It is an allocation: a plausible split of a known monthly total, guided by an annual sector shape and a monthly activity signal.

The sector mix moves with the seasons

The share chart shows the same numbers as a percentage of monthly demand. Buildings and city gas swell through the winter heating season, while industry and power take a larger share of the summer balance. Reading shares separately from levels makes it easier to see composition shifts that the stacked bars can hide.

The monthly shape comes from Carbon Monitor

Because no public dataset publishes monthly sector gas demand, the model borrows a monthly shape from Carbon Monitor China. Daily provincial CO₂ emissions are summed to national monthly totals for power, industry, residential buildings and ground transport. Each sector total is then turned into an activity index against its trailing 12-month average.

An index of 1.00 means normal activity for the time of year. An index of 1.20 means roughly 20% above the recent average. The model never converts emissions into gas volumes — Carbon Monitor is used only as a timing signal that says whether a sector was more or less active than usual in a given month.

How the allocation is built

Each sector starts from its 2023 IEA anchor share, is multiplied by its Carbon Monitor monthly index, and is then normalised across the four sectors. The normalised share is multiplied by JODI apparent demand for the month. So the four buckets always sum to the published JODI monthly total.

Two honest caveats are baked into the model. China’s stock change is reported as zero in this JODI extract, so apparent demand can include gas moving into or out of storage — actual end-use demand in those months may be lower (storage build) or higher (storage withdrawal) than the values shown. The power bucket is also a residual, not a clean gas-for-power estimate, because the public IEA page does not give one.

The 2023 annual anchor

The annual anchor fixes the long-run shape that Carbon Monitor then bends month by month. Industry, buildings and transport come straight from the public IEA China final-consumption table, converted from TJ gross to bcm with the IEA-reported Chinese heating value. Power is what is left when those three visible final-use buckets are subtracted from 2023 JODI apparent demand.

The source mapping is not one-to-one

Carbon Monitor and IEA do not share the same sector boundaries. Carbon Monitor has a residential category but no separate commercial-services proxy, so residential emissions shape the whole buildings bucket. Its power category shapes the broader power-and-residual bucket. Aviation is not used. The panel below shows exactly which source feeds which bucket, so the imperfections are visible and open to challenge.

Every modeled month, laid out

The table below shows the modeled sector volumes and shares for every month in the selected period, with the JODI apparent demand total in the second column. You can read across a row to see how a single month’s gas was allocated, or down a column to follow one sector through time.