Beijing's willows are now an air-quality decision, not a landscaping one

A direct emissions campaign at the Beijing Meteorological Tower found that biogenic sources made up roughly one-tenth of the city's measured VOC emissions but contributed nearly half of total VOC reactivity, the measure of how strongly emissions drive atmospheric chemistry ^[5]. Isoprene alone accounted for more than 90% of the chemical reactivity of those biogenic emissions ^[6], which moves tree species selection out of the landscaping budget and into the same ledger as solvent and fuel-vapor rules.

The setup matters because most urban BVOC numbers are model output. Direct measurements under real-world urban conditions remain rare, and current estimates rely largely on models built for natural ecosystems ^[3]. Researchers at Jinan University instrumented the 102-meter platform of the tower plus a laboratory at its base [c4a], and from May to July 2021 tracked chemical signals and turbulence fluctuations ten times per second [c4b], working with colleagues at the University of Innsbruck to apportion emissions among trees, traffic, chemical products, cooking and household activities [c4c].

Weight per unit emitted is where the asymmetry sits. On the reported shares, biogenic emissions carry roughly nine times more reactivity per unit emitted than anthropogenic ones ^[7]. Temperature then widens the gap: between 20 C and 35 C, the reactivity of VOCs from city trees rose seven- to eightfold, against about 1.4-fold for VOCs from human activity ^[8], so vegetation's share of total VOC reactivity climbed from 21% to 74% ^[9] and biogenic reactivity scaled with heat roughly five to six times faster than the anthropogenic term ^[10]. On days when ozone formation was particularly sensitive to VOC changes, observed peak ozone rose with temperature at nearly the same rate as VOC reactivity ^[11]. That is a conditional link, tied to a VOC-sensitive regime, not a general law.

The cross-city comparison is the operationally useful part. After adjusting for temperature and sunlight, Beijing's isoprene emissions were the highest among cities with direct measurements and approached levels found in temperate forests ^[12]. Vegetation cover did not explain why Beijing emitted several times more isoprene than cities with similar greenery ^[13]; comparing tree inventories and vegetation data across more than 20 cities pointed to species mix, with about 35% of Beijing's trees being isoprene emitters, including weeping willow and Chinese white poplar ^[14]. Across those cities, isoprene emission levels varied by more than tenfold and closely tracked the share of isoprene-emitting trees ^[15]. Several cities in Asia and Oceania may face comparable or greater pressure ^[16]. According to co-author Thomas Karl of the University of Innsbruck, European cities have considerably fewer isoprene-releasing species, with native trees often emitting monoterpenes, which also form ozone but are released in smaller amounts ^[17].

The background trend is that anthropogenic VOCs from solvents, paints, cleaning products and fuel vapors are increasingly regulated and measurements show them declining ^[2], which raises the relative weight of the biogenic term [c2b]. Karl says the findings do not argue for less greening or removing mature trees, but for adding BVOC emission potential to tree-selection criteria ^[18]. For Beijing, replacing high emitters in one-tenth of the urban tree population with low-emitting species during routine renewal could cut isoprene emissions by at least 29% ^[19] - a reduction close to three times the share of trees touched ^[20], which only works if the replacements are chosen from the top of the emissions distribution.

Watch whether any city publishes a species-level BVOC inventory alongside its canopy-cover target, and whether procurement lists for planting programs start carrying emission classes. The 29% figure is a scenario for one city, and the ozone link rests on VOC-sensitive days; both need replication elsewhere before they justify a planting rule.