Action 1: Epidemiological assessment of stomatal ozone flux-based critical levels for Southern European forests


Study area and sampling sites

The study was carried out in 2012-2013 in South-eastern France (30 plots) and North-western Italy (24 plots), where surface O3 levels are high and AOT40 levels are well above the threshold for forest protection in Europe, i.e. 5 ppm.h.

Figure1: The picture shows the area covered by the WRF-CHIMERE modelling system.
Black dots show the location of experimental plots.


Assessment of symptoms

At the 54 plots, 1080 trees were investigated over the two years. Assessments followed the ICP-Forests methodology and included three steps: symptom description, determination of causes, and quantification of the symptoms i.e.: (1) percentage of leaf/needle loss per tree (crown defoliation), (2) proportion of discolored leaves/needles per tree (crown discoloration) and (3) percentage of leaf/needle surface affected by O3 injury per leaf (visible foliar O3 injury). Symptoms were evaluated when the seasonal exposure to O3 and the probability of injury expression were highest, and before environmental conditions at the end of the growing season fostered the development of foliar senescence that can mask O3 injury. In South-eastern France and North-western Italy, these conditions occur at the end of August, when both campaigns were carried out. Every foliar injury assessment was done by a team of two trained observers, who were involved in the validation activities of the Expert panel on Ambient Air Quality of ICP-Forests, attended field courses organized by ICP-Forests and performed intercalibration exercises (GIEFS and IPLA).


Visible foliar ozone injury

A description of visible foliar O3 injury is available in the Deliverable D2 Refinement of criteria. Microscopy was used to verify the causes of visible injury on foliage and confirm visible foliar O3 injury as recommended by ICP-Forests (Figure2).

Figure2: Visible and microscopical ozone injury in two-year-old needles of Pinus cembra.
Chlorotic mottles, as discrete yellow patches with diffuse borders, occurred in irregular patterns,
usually close to stomata (A), cell collapse created large intercellular spaces
with anthocyans pigmentation (B, C). Circles mark areas close to stomata (B, C).


WRF-CHIMERE modelling system

Soil type and hourly meteorological (air temperature, relative humidity, soil water content and solar radiation), canopy moisture content (CMC) and O3 concentration data for 2012 and 2013 were obtained from the WRF-CHIMERE modelling system. For this work, data were provided at 1-h temporal resolution and 6x6 km spatial resolution.


AOT40 calculation

The O3 exposure index AOT40 was calculated as sum of the hourly exceedances above 40 ppb, for daylight hours (8am-8pm) during the assumed growing season according to the methodology for O3 risk assessment. In our study, the accumulation period started from the Start of the Growing Season (SGS) at a plot till the day when the survey of symptoms was carried out.


Phytotoxic Ozone Dose calculation

The stomatal O3 flux (Fst) was estimated using the DO3SE model as a function of the O3 concentration at the leaf boundary layer, the transfer of O3 across this boundary layer, stomatal conductance to O3 (gsto) and O3 deposition to the leaf cuticle. The stomatal O3 flux was then accumulated over a specified time interval and expressed as PODY i.e. Phytotoxic O3 Dose, where Y (nmolO3.m-2 PLA s-1) represents a detoxification threshold below which it is assumed that any O3 molecule absorbed by the plant will be detoxified (Mills et al., 2011). PODY was estimated for two thresholds Y: 1 nmolO3.m-2 PLA.s-1 as recommended by LRTAP Convention Mapping Manual (UNECE, 2010) and 0 nmolO3.m-2 PLA.s-1 as suggested by the fact that any O3 molecule entering into the leaf may induce a metabolic response.

PODY (mmol.m-2) was calculated as:

where PODY is the accumulated stomatal O3 flux above a threshold Y, calculated from hourly values of Fst, n denotes the number of hours to be included in the calculation period and dt = 1h (UNECE, 2010).

Both POD0 and POD1 were calculated over three time-windows during the growing season: A_PODY, from 8am to 8pm, a practical definition, adopted by the European legislation (Directive 2008/50/CE); B_PODY, using hours when cloud-free global radiation was > 50 W.m-2, as recommended by UNECE (2010); C_PODY, using hours with a global radiation > 0 W.m-2, assuming that stomata open even with < 50 W.m-2 light.

We used the available parameterizations for: continental central Europe deciduous broadleaf forests (Fagus sylvatica, Quercus cerris, Q. petraea, Fraxinus excelsior, Robinia pseudoacacia) at 19 plots, continental central Europe conifers (P. cembra, P. sylvestris, Abies alba and Picea abies) at 23 plots, and Mediterranean evergreen forests (P. halepensis) at 12 sites located along the coast.


Statistical analysis

Spearman test was carried out to understand the contribution of O3 (exposure or stomatal uptake) and meteorological and soil conditions to crown defoliation, crown discoloration and visible O3 injury. Random Forests Analysis was performed to determine the importance of each variable (annual O3 concentrations, POD0 and POD1, AOT40, canopy moisture content, temperature, global radiation, rainfall, relative humidity, soil water content) in determining visible foliar O3 injury, crown defoliation and crown discoloration.


Calculation of critical levels

We correlated PODY and AOT40 to real-world forest impacts in terms of different effect parameters, namely visible foliar O3 injury, crown discoloration and crown defoliation. New species-specific CLef were thus derived for each effect parameter, by joining data from all plots and years. All response functions selected to derive critical levels were statistically significant at p < 0.05.

As PODY was better correlated with visible foliar O3 injury than with defoliation and discoloration, and correlations were more significant for POD0 than for POD1, we selected visible foliar O3 injury as effect parameter and POD0 as ozone metric for defining PODY-based CLef values. Unfortunately, a definition of damaged tree/stand based on visible foliar O3 injury is missing in the literature. We thus based the selection of a visible foliar O3 injury threshold on a comparison of gas exchange of leaves with a range of visible O3 injury that was carried out in a 3-year-old O3-sensitive poplar plantation (Hoshika et al., 2012). CLef was thus derived from flux-effect functions for 15% of visible foliar O3 injury.


Occurrence of symptoms – Status of health forest

Out of 54, 28 plots in 2012 and 32 plots in 2013 showed mean crown defoliation exceeding 25% (data not shown). Average defoliation and discoloration were higher in conifers than in broadleaves. Defoliation and discoloration were higher in F. sylvatica than in F. excelsior and R. pseudoacacia. P. halepensis showed higher mean defoliation and discoloration than P. cembra and P. sylvestris.

Out of 54, 16 plots in 2012 and 18 plots in 2013 showed mean percentage of foliar surface affected by visible O3 injury exceeding 15% (data not shown). Visible foliar O3 injury was higher in F. excelsior than in F. sylvatica and R. pseudoacacia. Quercus species (Q. cerris and Q. petraea) showed very low levels of visible foliar O3 injury (around 0.3%). In conifers, chlorotic mottling was missing in current-year needles, and higher in C+2 needles than in C+1 needles. Visible foliar O3 injury was higher in P. cembra than in P. halepensis in contrast with the response of crown defoliation and discoloration. P. sylvestris was less affected by visible foliar O3 injury (<10%), and P. abies and A. alba were not injured (not shown).

Ozone produces a very characteristic chlorotic mottling on the needles of P. halepensis and P. cembra, therefore, both species are valuable bioindicator. Our results allowed ranking the tree species on the basis of their sensitivity to ambient O3. P. halepensis and P. cembra were moderately and highly O3-sensitive species, respectively. Pinus sylvestris was less impacted by O3 visible injury, and can be considered as a low O3-sensitivity species. Abies alba and P. abies were not impacted and may be classified as a O3-tolerant species. For broadleaved species, F. sylvatica and R. pseudoacacia can be considered as moderately O3-sensitive species and F. excelsior as highly O3-sensitive species. Quercus species (i.e. Q. cerris and Q. petraea) showed very low levels of visible foliar O3 injury and may be classified as an O3-tolerant species.


Discussion, conclusions and recommendations

Forests have many important functions for economic activity, nature conservation, environmental protection, and carbon sinks. The Mediterranean area is sensitive to climate change because it represents a transition zone between arid and humid regions of the. Over Southern Europe, the summers will be increasingly characterized by warm, dry weather with calm winds (Sicard and Dalstein-Richier, 2015), favoring high O3 levels and drought. In addition, Southern Europe is representative of water-limited environments that cover about 41% of Earth's land surface.

To protect forests against O3 pollution, appropriate standards and realistic thresholds, representative of actual field conditions, are needed. AOT40 is the current standard for the protection of vegetation in Europe. PODY is under discussion as new standard but a validation of the standard and of the threshold Y under field conditions is still missing. A standard for forest protection is biologically relevant when it translates into real-world forest impacts. Until now, however, the definition of flux-response functions in Southern Europe has been limited to simulated experiments. In this innovative epidemiological assessment of forest responses to O3, we compared the performance of the three O3 indices (AOT40, POD1 and POD0) used in Europe as descriptors of O3 risk for vegetation. We showed that PODY was better correlated with visible foliar O3 injury whereas AOT40 was better correlated with crown discoloration and defoliation, i.e. typical aspecific indicators caused by multiple factors (e.g. tree species, forest management, meteorology, site and soil characteristics, water limitation).

This study confirms that the responses of forests to O3 depend not only on atmospheric O3 concentrations but also on the O3 uptake through stomata (stomatal flux) into the leaves. We thus conclude that PODY is better than AOT40 as a metric of O3 risk for forests in Southern Europe.

Better flux-response relationships with visible O3 injury were obtained when there was no threshold (POD0) above which flux was accumulated than when POD1 was used. The results suggest that any O3 molecule entering into the leaf induces a metabolic response. In addition, the sensitivity of PODY to the inputs typically increases with the value of the threshold Y.

We thus recommend keeping the value Y as low as possible (POD0) as the most biologically-based threshold.

Following a practical definition, PODY is accumulated from 8am to 8pm or for the daylight hours with global radiation > 50 W.m-2. We showed that this implied an underestimation of the real O3 uptake.

As stomata cannot be considered as totally closed during the above time-windows, we recommend applying a 24-h time-window for POD accumulation, with a non-null global radiation because it has both biological significance and practicality in usage.

Our study highlighted that the most important environmental variable affecting visible O3 injury, in all tree species, was soil water content while the annual O3 concentrations had the lowest relative importance. The DO3SE model must thus include the SWC function, because it is critical for water-limited environments.

For the first time, an epidemiologically-based toxicology study was carried out here, linear flux-response relationships were established, and correlated with real plant symptoms assessed under natural environmental conditions.

As a main conclusion, we recommend the following generic flux-based critical levels CLef: 19 mmol.m-2 for highly O3-sensitive conifers (i.e. P. cembra); 19 mmol.m-2 for broadleaved species (i.e. Fraxinus excelsior); 24 mmol.m-2 for moderately O3-sensitive conifers (i.e. P. halepensis) and 21 mmol.m-2 for moderately O3-sensitive broadleaved species (i.e. Fagus sylvatica).

This innovative study provided useful information for establishing the best standards and thresholds for forest protection to O3. Derivation of the new flux-based critical levels for Southern European tree species represents a considerable progress in the development of methods for quantifying effects of O3 on vegetation at the regional scale. These results will serve as a decision-support tool for European authorities. With information in hand, policymakers can make informed decisions about proposed changes to legislation to scientifically assess the effectiveness of air pollution control strategies in European forests.