Biogenic isoprene emissions, dry deposition velocity and surface ozone concentration during summer droughts, heatwaves and normal conditions in Southwestern Europe

. At high concentration, tropospheric ozone ( O 3 ) deteriorates air quality, inducing adverse effects on human and ecosystem health. Meteorological conditions are key to understand the variability of O 3 concentration, especially during extreme weather events. In addition to modifying photochemistry and atmospheric transport, droughts and heatwaves affect the state of vegetation and thus the biosphere-troposphere interactions that control atmospheric chemistry, namely biogenic emissions of precursors and gas dry deposition. A major source of uncertainty and inaccuracy in the simulation of surface O 3 during 5 droughts and heatwaves is the poor representation of such interactions. This publication aims at quantifying the isolated and combined impacts of both extremes on biogenic isoprene ( C 5 H 8 ) emissions, O 3 dry deposition and surface O 3 in Southwestern Europe. First, the sensitivity of biogenic C 5 H 8 emissions, O 3 dry deposition and surface O 3 to two speciﬁc effects of droughts, the decrease in soil moisture and in biomass, is analyzed for the extremely dry summer 2012 using the biogenic emission model


Supplement of
Biogenic isoprene emissions, dry deposition velocity, and surface ozone concentration during summer droughts, heatwaves, and normal conditions in southwestern Europe Antoine Guion et al.
Correspondence to: Antoine Guion (antoine.guion@ineris.fr) The copyright of individual parts of the supplement might differ from the article licence.

Figure S3 .
Figure S3.Mean ∆LAI [m 2 /m 2 ] between summer (JJA) 2012 and 2014 (upper left), relative difference of LAI (lower left) and time series of LAI spatially averaged over the Southwestern Europe (lower right).

Figure S4 .
Figure S4.Validation scores of the daily maximum surface O3 [µg/m 3 ] for summer (JJA) 2012, 2013 and 2014 ("Reference" simulations), with the European surface network observations AQ e-Reporting.Left column: mean bias.Right column: temporal correlation (R pearson).The number of stations taken into account is 167, 188 and 207 respectively for summer 2012, 2013 and 2014

Figure S6 .
Figure S6.Daily mean 2m temperature, shortwave radiation and soil wetness simulated by the WRF model during summer 2012 over the Balkans (upper panel), Pô Valley (middle) and Central Spain (lower).

Figure S7 .
Figure S7.Daily mean chemistry regime parameter [α] averaged over the summer 2012, 2013 and 2014 ("Reference" simulations).α calculates the ratio of the reaction rate of RO2 radicals with NO (high-NOx regime) with respect to the sum of reaction rates of the reactions with HO2 and RO2 (low-NOx regime).It gives a relative indication of low-N Ox (low α, about 0.5) and high-N Ox (high α, about 0.9) regime areas that are detailed in Zhang et al. (2013).More information about the calculation method of α is provided on the online documentation (https://www.lmd.polytechnique.fr/chimere/).

Figure S8 .
Figure S8.Daily mean O3 surface concentration [µg/m 3 ] during summer 2012, spatially averaged over the Balkans (upper panel) and Central Spain (lower panel) from the EEA observations and the different CHIMERE experiments.

Figure S9 .
Figure S9.Simulated weather conditions (2m temperature, shortwave radiation, cloud fraction and soil wetness) by the WRF model over the Southwestern Europe, clustered by identified extreme weather events (from the RegIPSL model).The number of days is in parentheses.The analyzed period is summer 2012, 2013 and 2014, covering a total number of 276 days.

Figure S10 .
Figure S10.Land cover fraction of cropland, grassland and forests over the Southwestern Europe from USGS (left column) and MODIS MCD12 product (right column).

Figure S11 .
Figure S11.Daily HCHO total column [molecules/cm 2 ] during summer 2012 observed by OMI (OMHCHOd level 3 product) and simulated by CHIMERE, averaged over the Balkans (upper left panel), Pô Valley (upper right panel) and Central Spain (lower panel).A moving average window of 3 days is applied on observations and only days with at least 30% of data over the spatial cover are kept.The number of observations (OMI pixels) is indicated below each panel.