Outdated East Asia Forest Maps Distort Air Pollution Forecasts

Somewhere within the code used for numerous air quality forecasts is a record of East Asia’s forests preserved as they were in 2003. Facebook had not yet been created, and George W. Bush was serving his first term as US president.

The landscapes shown in those ageing satellite images bear little resemblance to the vegetation that is actually present today.

New research suggests this outdated information is subtly distorting forecasts for ozone and particulate pollution, since plants are not merely part of the backdrop.

As vegetation adapts to a warming climate, it alters the quantities of pollution-producing gases released into the atmosphere. However, the models used to monitor these emissions have not been updated accordingly.

The research was conducted by Hyo-Jung Lee and Yu-Jin Jo of Pusan National University in South Korea, with Younha Kim of the International Institute for Applied Systems Analysis in Austria.

Forest cover that has since disappeared

Plants emit biogenic volatile organic compounds, particularly isoprene and monoterpenes, which react in the atmosphere and contribute to the formation of ground-level ozone and biogenic secondary organic aerosols.

Both pollutants affect the air that people breathe. The amount of these compounds emitted by an area depends on the type and extent of vegetation growing there.

A shift in forest cover is not simply a small ecological detail; it directly affects how polluted the air can become.

That is the central issue: many atmospheric chemistry models still rely on vegetation information dating from 2003.

Since that time, East Asia’s ecosystems have altered substantially as a result of reforestation, land-use changes and wider climate-warming effects.

In practical terms, these models are simulating forests that no longer exist, reducing the accuracy of air pollution forecasts.

Testing updated vegetation maps

To assess the scale of the effect, the researchers used the WRF-Chem atmospheric chemistry model alongside MEGAN, a widely used module for biogenic emissions.

They substituted its standard 2003 vegetation dataset with satellite observations from 2024.

All other factors, including the meteorological conditions used in the simulation, were kept unchanged. As a result, any difference in the results could be attributed solely to vegetation.

“Our results show that updating vegetation information alone can substantially change biogenic emissions, O₃, and biogenic secondary organic aerosols (BSOA),” Lee said.

Different forests produce different results

The satellite observations revealed a more complex picture than a uniform increase in greenery. Vegetation had generally become denser across East Asia since 2003, although the pattern varied considerably by location.

Some parts of China became markedly greener, whereas vegetation declined in parts of Japan. Once the researchers entered this uneven landscape into the model, predicted ozone and BSOA levels showed similarly uneven changes.

Suburban locations experienced the greatest rise in pollutant formation. These areas occupied a favourable combination of vegetation change and nitrogen oxide concentrations, leading to more secondary pollution.

Urban areas, while already containing abundant nitrogen oxides, showed little change because roads and buildings left limited space for expanding vegetation.

Rural locations displayed the reverse trend. Although they gained substantial vegetation, nitrogen oxide concentrations were too low for that growth to generate much extra pollution.

Put simply, biogenic emissions represent only one half of the process. The substances with which they react are equally important.

The impact of obsolete forest maps

This variation explains why vegetation data that is decades old matters more than might initially be expected.

A model using outdated inputs is not merely slightly inaccurate across all areas. It is likely to overlook significant local differences in the way pollution develops. This is important when the aim is to alert a specific city to a day of poor air quality.

“Our findings demonstrate that regularly updating vegetation datasets in biogenic emission modules can improve the accuracy of atmospheric chemistry models,” Lee said.

The researchers said the results underline the need to update vegetation datasets on a routine basis.

This would enable atmospheric chemistry models to represent evolving ecosystems and their interactions with the atmosphere more accurately.

Updating air pollution models as forests evolve

The team considers that incorporating current satellite observations into air pollution models could improve daily air quality forecasts and provide policymakers with better information as conditions continue to evolve.

The analysis examined August 2024, as vegetation activity and biogenic emissions reach their highest levels during this period of the year.

Future research will use the same method for other seasons and longer timeframes. This should offer a clearer understanding of how an ever-changing landscape alters the atmosphere above it.