Autumn leaf colouring became later at East German stations with high early-1990s SO2 after the SO2 fall, but the delay was small, faded after 2010 and was uncertain in a permutation test
Question: after the SO2 emissions of former East Germany fell by more than 80% in the 1990s, did tree phenology change more at places that had high SO2? Data: DWD phenology 1961-2023 (93,217 autumn-colouring records at 1,580 East German stations; leaf unfolding and spring flowering as comparison classes); early-1990s SO2 interpolated from 196 EEA AirBase monitors (38 in the East; station median 32 ug/m3); E-OBS weather. Design: a plan written before the models; station x species and year x species fixed effects, so only differences inside the East are used; weather terms; change of the exposure slope from 1976-1990 to 1996-2008; errors clustered on 81 spatial blocks. Result: autumn colouring of birch, oak and horse chestnut became later by +0.37 days per 10 ug/m3 (95% CI +0.03 to +0.71), about 2 days across the East German range. The pre-trend was +0.09 (CI -0.07 to +0.25). Leaf unfolding did not change (-0.03) and spring flowering became earlier (-0.18). The triple difference against flowering was +0.55 (CI +0.21 to +0.89). Thus the planned decision rule was met. But the delay was largest in 2001-2010 and near zero after 2015. A post-hoc block-permutation test gave p = 0.16 (p = 0.04 with latitude and altitude terms). Several variants had CIs that included zero. West Germany had too small an SO2 gradient to replicate the test. Relative to West Germany, all East stations shifted after 1995 (autumn +1.4 days, spring -1 to -1.9 days), but this mixes the SO2 fall, warming and the 1991-1992 change of the observer network. Limitations: few East monitors before 1994; no pre-1990 exposure data; no observer identity; confounding by the economic change in the polluted south. The evidence agrees with a small temporary recovery effect but does not establish an SO2 cause.
air pollution · sulfur dioxide · SO2 · plant phenology · autumn leaf colouring · leaf senescence · natural experiment · German reunification · East Germany · DWD phenology · EEA AirBase · E-OBS · difference in differences · event study · observation bias · observer network change · negative control · permutation test · climate change · pollution recovery
References
7- useshttps://opendata.dwd.de/climate_environment/CDC/observations_germany/phenology/annual_reporters/wild/historical/
Phenology dates (DWD annual reporters, wild plants).
- useshttps://eeadmz1-downloads-api-appservice.azurewebsites.net/
EEA air quality download service, dataset 3 (AirBase) SO2 for Germany.
- useshttps://discomap.eea.europa.eu/map/fme/metadata/PanEuropean_metadata.csv
Monitor coordinates and station types.
- usesdoi:10.1029/2017JD028200
E-OBS gridded daily temperature and precipitation (v31.0e).
- reviewshttps://ceip.at/the-emep-grid/gridded-emissions
We checked the 1990 gridded SOx; it does not resolve the GDR sources, so we did not use it.
- usesDrier summers were followed by later, not earlier, autumn leaf colouring at German phenology stations in 1951-2024; the Swiss estimate had the same sign but was uncertain sci:sha256:70b69f10765f3c0c6114e4131551020643445032d887563a907e07979caab796
Same DWD network; we reuse its fixed-effects code (fe.py) and screening rules, and extend the autumn-colouring analysis to air pollution.
- Soil moisture increased the ozone injury response of bioindicator plants at the same eastern US forest biosites in 1994-2010, but ozone alone did not predict year-to-year injury sci:sha256:268004a08eb37a3a36207332736b005b07fe9d13a4687a19b1e966f96044ecb6
Related campaign test of ozone injury in bioindicator plants.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.