Monitoring canopy exposure to ozone

As a result of vegetation uptake and interactions which other surfaces, ozone concentrations vary with height, particularly above and within forests as a result of their high capacity for ozone uptake. Ozone concentrations are generally measured at standard heights (between 2 and 4 m) in open areas, and thus significant errors can be introduced in estimating canopy exposure.

Technical details

Passive samplers located above the canopy on a meteorological mast Passive sampler exposure assembly

At a number of the Level Network II sites, ozone concentrations are measured above the canopy using passive samplers. These data are compared with concentrations measured at standard measurement height (2 m) beneath the canopy and also at a nearby open area. In the latter case, passive sampler data are also compared with concentrations measured by a co-located active, chemiluminescence monitor, allowing exposure to be calculated as AOT40.  Radiation (and rainfall) shields protected all samplers. Duplicate diffusion tubes were changed during routine site visits for deposition collection (every two weeks), with samplers transported in thermally insulated boxes. For each two week period, one blank was retained in a fridge at Alice Holt, whilst a second blank was transported with the analysis tubes to a remote site, and installed beneath the radiation shield within its protective tube. Samplers were forwarded to the analysis laboratory within 48 hours of collection, with analysis carried out within 10 days of collection. The active samplers were located within temperature controlled mobile units (20-30oC) with data recorded as 30 minute means of 10 second measurements using data-loggers.


The relationships vary greatly between sites, as do the implications for uptake or flux modelling. Data are shown only for Alice Holt and Grizedale, although the measurement programme is on-going, with data now available for Thetford (East Anglia) and Coalburn (Kielder forest, Northumberland).

Graph - ozone concentration - Alice Holt

Graph - ozone concentration - Grizedale

Concentrations recorded below the canopy are generally lower than those above the canopy, indicating some canopy uptake. In general, concentrations at the forest plot are also slightly higher than at the open plot which may be due to (i) small altitudinal differences; (ii) topographical effects; (iii) passive samplers being more efficient above the canopy because of the higher wind speeds and turbulence, or a combination of effects. There is some evidence in these data-sets that the above canopy: open plot differential is at a maximum at low windspeeds. Indeed, it is known that diffusion samplers are inefficient at windspeeds below about 0.5 m s-1, and further ventilation is suggested where this is likely to occur. Standard height (~2 m) monitoring networks using passive samplers may therefore underestimate ozone concentrations if ventilation fans are not installed.

The table below provides an estimate of AOT40 above the canopy by scaling the concentrations measured by the active sampler according to the concentration difference between the open plot and above canopy samplers. These results indicate that the ozone concentrations to which forest canopies are subjected may be underestimated and, therefore, that ozone-induced damage may be potentially far more serious than previously thought, since injury thresholds have largely been based on chamber studies in which ‘canopy’ level concentrations are measured. In the case of Grizedale, the calculated AOT40 of 21.6 ppm h brings canopy exposure well above the critical level of 10 ppm h.

Summary of the ozone pollution environment at the two Level II sites from 10 April 2000 to 31 October 2000 as measured by the active monitors at each site.
(Figures in parentheses represent above canopy concentrations extrapolated from the difference between concentrations measured by passive samplers at the open plot and above the canopy)
Site AOT40 - ppm h 7 hour mean - ppb 24 hour mean - ppb
Alice Holt 3.56 (5.69) 29.6 (32.9) 22.1 (24.5)
The Lakes 6.07 (21.7) 33.8 (42.8) 27.1 (34.3)

At Alice Holt, ozone concentrations were generally lower than at Grizedale resulting in lower AOT40 (3.6 vs 6.1 ppm h), 7 hour mean (30.0 vs 33.8 ppb) and 24 hour mean values (22.3 vs 27.1 ppb). These lower values are likely to be a result of night-time concentrations generally falling to lower values at Alice Holt as a result of more stable atmospheric conditions and higher rates of NO2/NO/O3 inter-conversion. These lower night-time concentrations may also explain the larger discrepancy between active and passive monitoring concentrations at Alice Holt as compared to the Lakes.


Data collected using active and passive samplers generally showed good agreement at both sites, although passive samplers did appear to overestimate concentrations at Alice Holt, where low night-time concentrations were commonplace as a result of the NO/NO2/O3 cycle, and stable atmospheric conditions. In addition, it is apparent that passive samplers may not necessarily give a indication of ozone concentrations under varying conditions of windspeed. This is highly pertinent to studies assessing the vertical distribution of ozone concentration through a forest canopy. Notwithstanding these areas of uncertainty, it is evident that large vertical gradients in ozone concentration do exist, and that the magnitude of this gradient is dependent to a large extent on the canopy structure, and atmospheric stability, which is, itself, a function of topography. When the impact of ozone on ground vegetation is assessed, any vertical gradient should be assessed. This is clear for Alice Holt, where ozone concentrations were minimal beneath the canopy for a large portion of the growing season.

The data also suggest that the magnitude of local scale variations in ozone concentration may be significant when impacts of ozone on specific sites are investigated. Crucially, the relatively small differences in mean concentration between the open plot, and above the canopy, as measured by passive samplers, can translate to large differences in AOT40, and as a consequence, to potentially significant differences in predictions of ozone mediated damage. This point is particularly important for critical load exceedance calculations based on the AOT40 exposure index, and provides further justification for the development of flux-based critical loads.