C.
Nali
*,
A.
Francini
and
G.
Lorenzini
Dipartimento di Coltivazione e Difesa delle Specie Legnose “Giovanni Scaramuzzi”, Università di Pisa, Via del Borghetto, 80, 56124, Pisa, Italy. E-mail: cnali@agr.unipi.it; Tel: +39 050 960092
First published on 24th November 2005
Tropospheric ozone is a growing environmental menace in Italy and in the whole Mediterranean basin. The importance of active biomonitoring of this pollutant with hypersensitive Bel-W3 tobacco plants is stressed, and several examples of field studies carried out in Italy with this technique are presented. Current limitations are discussed, with special emphasis on data quality assessment and the opportunity of adopting easy-to-use kits based on tobacco germlings instead of adult plants. A standardization of methodologies (from cultivation to scoring and data elaboration), also at an international level, is strongly felt to be needed, in order to get official acknowledgement of biomonitoring procedures. Potential educational implications, with the active involvement of students and environmentalists, are shown. Other biological indicators are used, namely sensitive and resistant white clover (Trifolium repens) clones (as descriptors of biomass reduction in crops species) and Centaurea jacea (brown knapweed) as a model species to evaluate the relationship between ozone exposure and effects on the performance and injury symptoms of native plants which are largely used in the framework of European programmes.
![]() | Cristina Nali was born in Italy, in 1965. She received her PhD in Phytotoxicology from University of Pisa, Italy in 1997. During this time and thereafter she worked as a researcher on plant pathology, with special concern about the interactions between fungal diseases and gaseous air pollutants. Throughout this time, she also engaged in biological education projects, with particular regard to biological monitoring of ground-level ozone; another major field of interest has been the study of the effects of environmental pollutants upon qualitative performances of agricultural and medicinal plant species. Currently, she is Assistant Professor at the Faculty of Agricultural Sciences of the University of Pisa. Her research interests are interactions between air pollutants and plants, especially biological monitoring and the study of physiological and molecular bases for differential response of plants to the oxidative stress. |
The on-ground ozone distribution in Italy is a relatively young field of research: in this country, measurements of ambient ozone are mainly made using instruments operating on the principle of UV absorption. Some 50 analyzers were regularly operating about ten years ago:4 unfortunately, their spatial distribution was irregular and large geographical areas—especially in the south—were not covered. For instance, in Tuscany (central Italy, 22 997 km2, about 3.5 × 106 inhabitants, 287 municipalities), ozone analyzers operate routinely in only 13 municipalities, covering merely 6% of the total surface area, 33% of the inhabitants and 4% of the municipalities. Moreover, only a handful of these sites have long-term records. Information is scattered and no nationwide database has been set up so far.
Evidence of ozone-induced visible phytotoxic markings was first reported in Tuscany in the early 1980s,5 but the proof of detrimental effects of regional air pollution on plant productivity under field conditions was only obtained several years later. The establishment of air filtration experiments in northern Italy, based on the open-top chamber exposure technique,6,7 showed the relevant beneficial effect of filtering ambient air on productive performances of several crops. In bean plants8 the appearance of typical interveinal brown stipple associated with mesophyll histopathological alterations—which were reproduced by fumigating healthy plants with realistic doses of ozone—was positively attributed to ozone toxicity. Specific symptoms were observed on the leaves of peach trees grown in open-top chambers at a rural site in central Italy.9,10 reported ozone-induced visible injury in Italy on several species, such as bean, courgette, grapevine, peach, pepper, radish, soybean, spinach and tomato. Symptoms attributed to ozone have also been reported for natural vegetation: black locust, heaven tree and poplar,11 as well as for conifers growing in urban areas.12
Comparison of the observed ozone concentrations with United States’ crop damage criteria suggests that in summers the quantitative yield of many crops will be reduced in a significant way. Potential ozone damage in connection with crop susceptibility has been evaluated for the Province of Milan via a Geographical Information System (GIS) approach.13
Ozone exposure at rural and forest sites in Italy exceeds by far concentration-based critical levels.14 The computation of the AOT40 (accumulated exposure over a threshold of 40 ppb) values for the area of Pisa (Tuscany)15 (Fig. 1) has systematically indicated values up to 20 ppm h (for three months, June–August, daylight hours): these figures are well above the long-term critical level for agricultural crops (3 ppm h for three months).16 Similar results have been reported from other Italian regions, such as Campania.17 A common trait is that daily patterns of ozone concentrations in hilly areas are different from those of sites located in plains, the differences between daily maximum and minimum being quite low at the hilly sites.
![]() | ||
Fig. 1 Cumulate AOT40s of ozone during the photochemical season 2004 in the Tuscan sites of Lari (rural, 130 m above sea level), Montecerboli (remote, 370 m a.s.l.) and Pisa (urban, 5 m a.s.l.); the bold lines indicate the thresholds for critical levels of 3 ppm h (for three consecutive months, the target to be protected being quantitative yield of agricultural crops) and 10 ppm h (for six consecutive months, the target to be protected being biomass reduction of forest trees).16 |
The estimated yield losses for soybean due to ozone pollution in some Tuscan districts ranged from 17 to 27% of the potential in ozone-free air.18
An assessment of the biological implications of air pollution should improve global knowledge of environmental contamination also in economic terms and contribute to urge authorities and decision-makers to put into action proper measures to improve the quality of our atmosphere. It is interesting to note that the first perception of an air pollution problem in the Los Angeles area emerged from the observation of ‘unusual’ plant injury in the late 1940s.19 Indicator plants can be successfully used to detect the presence of phytotoxic ambient levels of ozone over large territories, integrating the data from analyzers.
Suitable bioindicators may be useful in involving ordinary people in environmental diagnosis. One of the key aspects of sustainable development is the active participation of citizens in environmental issues, such as air quality monitoring.20 However, the final goal of biomonitoring is not to replace the conventional physico-chemical approach: an integration of the two systems is the most appropriate solution for air quality assessment.
‘Conventional’ biomonitoring is performed with adult tobacco plants (about 2 months old, 60 cm in height). Important operational constraints are to be stressed: (i) a large space (and a long time) in filtered-air, controlled-environment growing facilities is required; (ii) due to leaf fragility, long-distance transport of plants is difficult; (iii) visual assessment of leaf injury should be performed by remote operators, all of whom must be adequately selected and skilled; (iv) indicator plants stay in a site for many (usually 4) weeks, and their sensitivity may change over time (as a rule, during the first days of exposure in the open-air the plants are overresponsive).22 A preliminary attempt to standardise the procedure of tobacco exposure has been carried out in Germany.23
In all these surveys, hypersensitive tobacco Bel-W3 plants always showed visible injury, demonstrating that the threshold concentration of ozone required for the appearance of visible symptoms (40–50 ppb for exposures of about 5 h, or 30 ppb for 8 h21) had been largely exceeded. Weekly variation in the amount of leaf injury was constantly reported and generally was associated with meteoclimatic parameters. A linear multiple correlation model has shown that leaf damage on indicator plants in Milan was influenced positively by temperature and negatively by vapour pressure deficit.28 A campaign involving many stations29 showed, through a chronological analysis, a high correlation between all the sites in the Po Valley, even for those hundreds of kilometres distant from each other. This can be explained by the existence of synchronous patterns of ozone pollution on a wide geographical scale.
Integrated ambient air studies have been conducted in the city of Florence.30,31 The results obtained with biomonitoring were in agreement with those offered by instrumental monitoring; as expected, the city centre has the lowest leaf injury values. This pattern is consistent with what it is expected for ozone in large urban areas, as ozone depletion by nitrogen oxide is higher in the most polluted environments.32 In the same way, the amount of visible injury on the indicator plants was negatively correlated with the presence of lead in/on the leaf tissues of ryegrass (accumulator plant for heavy metals). The reason for this is that most ambient lead came at that time from vehicle exhausts, which also contain a mass of other pollutants, which easily react with ozone molecules and deplete them.
Some years ago, an innovative procedure for ozone biomonitoring was developed and patented at the University of Pisa.5 Tobacco seeds are germinated in an ozone-free environment and when 15 days old (the first leaf 1 cm long), the germlings are singly transplanted in commercial polystyrene tissue culture plates (13 × 9 cm) with 24 round wells (16 mm diameter, 20 mm depth), filled with an organic compost (Fig. 2). The plates are modified to allow the watering of the wells. After a further day in ozone-free air for acclimatization, the plates are exposed to ambient air. A typical configuration is with 16 seedlings of the hypersensitive Bel-W3 tobacco and 7 of the ozone-resistant Bel-B. The 24th well is connected to the mouth of an inverted feeding bottle containing deionized water, to provide continuous sub-irrigation of the seedlings. At least 3 plates are routinely positioned in each monitoring site under a metal cage covered with a fine shading net. Shading is a normal practice in biomonitoring with tobacco, as the cultivars utilized are shade-grown. These plates are exposed to ambient air for 7 days. The intensity of the injury of cotyledons and the first leaf is visually assessed and related to the dose of ozone to which the seedlings have been exposed. The necrotic area percentage was estimated by comparison with standards and reported in terms of Cotyledonar (CII) and Leaf Injury Index (LII), based on a 1 to 5 scale (1: no lesion; 5: more than 50% of surface covered by necrosis). Concerning the pathometric scale, every plant pathologist is aware of the crucial role of the Horsfall–Barratt (H–B) rating scale for measuring rust and other fungal diseases, but in our opinion this is not the right tool to measure the extent of foliar injury induced by ozone. This is due to the fact that the H–B scale punctually covers all the span from 0 to 100% damaged surface; since long we know that ozone induced lesions may spontaneously coalesce causing the collapse of large regions of mesophyll, even several hours after the end of the exposure. This implies the inopportunity of rating large necrotic surfaces, so it is common in the field of ozone monitoring with tobacco to ignore injury above about 40% of the whole surface (see ref. 27 for a detailed analysis of the matter).
![]() | ||
Fig. 2 A kit of tobacco Bel-W3 germlings after seven days of exposure to ambient air. |
The advantages of ozone biomonitoring with this miniaturized kit may be summarized in the following way: (i) only a very limited space is required in the growing facility; (ii) a large number of individuals may be exposed; (iii) the biomonitoring units may be easily dispatched to long distances and returned to the main laboratory for visual assessment of injury; this ensures a standardization of the reading; (iv) the permanence of germlings for a standard time (e.g. 24 h) in an ozone-free environment allows the full maturation of symptoms due to the last-day exposure.
This innovative methodology is reliable and allows an inexpensive survey of the geographical distribution of tropospheric ozone over large areas, including remote sites, where conventional monitoring would be difficult. A large scale campaign with the use of the miniaturized kit has been performed in seven districts of Tuscany.33 Another trial demonstrated the presence of phytotoxic ozone in the small remote islands of Capraia and Gorgona (Tyrrhenian Sea), where local sources of precursors are negligible.34 Assessments with this kit were also carried out in France and the UK.35,36
Epiphytic lichen biodiversity distribution has been widely used as a bioindicator of air pollution throughout the world, mainly due to their high sensitivity to sulfur dioxide.37 Comparative evaluations of ozone hypersensitive tobacco responses and lichen biodiversity in several locations indicated that the two variables investigated did not show any degree of association.38
As visual estimates of foliar injury involve subjectivity, there are limitations to the tobacco systems. A methodology based on non-subjective measures, rather than estimated responses to ozone, and not significantly influenced by other factors would clearly be an addition to, and an improvement over, present systems. Ozone-sensitive (NC-S) and ozone-resistant (NC-R) clones of white clover (Trifolium repens L.) cv. Regal have been tested in the USA to determine this feasibility to indicate concentrations of ozone and its potential effects on plants.39 The clones display measurable differences in growth response to ozone while responding similarly to the other factors that influence plant growth. The growth of NC-S and NC-R is similar when ozone concentration is low, whereas the NC-S grows much less than NC-R when ozone concentration is high (Fig. 3). A comparison of this method with the classic exposure in open-top chambers (filtered vs. unfiltered air) conducted by an Italian group confirmed its suitability.40 The utility of this method has been acknowledged by workers who carry out research on effects of air pollution on crops and non-woody plants within the framework of a UN/ECE International Cooperative Programme (ICP-Vegetation, formerly ICP-Crops41) for the Convention on Long-Range Transboundary Air Pollution. Within this programme,42 clover experiments are performed in several Italian locations. Data so far available in Italy show no clear relationship between actual AOT40 and yield losses of white clover: in experiments in Milan, Naples and Rome (1996–2001), these were not significantly correlated with their respective AOT40 (n = 22; R2 = 0.08), whereas they did correlate in a survey carried out in Pisa in 1999–2000.15 The same survey, using white clover NC-S/NC-R epigeous biomass ratio as response indicator showed that AOT60 and AOT80 yield higher R2 than AOT10, AOT20, AOT30 and AOT40 (Table 1).15 Our determination coefficients were much higher than those reported for the European and Mediterranean scenarios,43–45 and slightly lower than those reported in the USA.46Table 2 reports the statistical analysis performed with the combined O3 (AOT40 accumulated over three months) and biomass data, omitting the data of the last two harvests, to remove any late season climatic effects on plant biology. The extent of biomass reduction was closely related to the O3 dose received over the previous three months (June, July and August). This is in accordance with ref. 47 whose authors hypothesized a residual carry-over effect in the biomass responses of NC-S/NC-R between subsequent harvests in a given growth season. Fig. 4 describes the typical response of the NC-S/NC-R biomass ratio. It should be stressed that this experimental approach can help to assess the economic losses caused by the effects of ozone on crops. A modified approach, based on the use of clover ‘mini-stations’ has been successfully developed by Roman scientists.48 AOT40 descriptor were also positively correlated with the visible injury of hypersensitive tobacco cv. Bel-W3 (n = 8; R2 = 0.75).38
![]() | ||
Fig. 3 Biomass production of white clover (Trifolium repens) cv. Regal NC-S clone (left) compared to that of NC-R clone (right) in a monitoring campaign concerning tropospheric ozone conducted in Pisa. |
![]() | ||
Fig. 4 Response of the NC-S/NC-R biomass ratio of white clover to cumulated AOT40 over 28 monthly harvests between 1997 and 2004 in Pisa; broken lines represent the confidence interval (P = 0.05) (y = 0.928–2.830 e−5x; n = 28; R2 = 0.66). |
Year | Regressors | Equation | R 2 |
---|---|---|---|
1999 | AOT10 vs. NC-S/NC-R | Y = −9.08 X + 1.06 | 0.41 |
AOT20 vs. NC-S/NC-R | Y = −0.02 X + 1.10 | 0.45 | |
AOT30 vs. NC-S/NC-R | Y = −0.03 X + 1.14 | 0.50 | |
AOT40 vs. NC-S/NC-R | Y = −0.06 X + 1.20 | 0.55 | |
AOT60 vs. NC-S/NC-R | Y = −20.64 X + 61.86 | 0.64 | |
AOT80 vs. NC-S/NC-R | Y = −448.68 X + 63.11 | 0.49 | |
2000 | AOT0 vs. NC-S/NC-R | Y = −1.69 X + 0.71 | 0.55 |
AOT10 vs. NC-S/NC-R | Y = −2.30 X + 0.72 | 0.57 | |
AOT20 vs. NC-S/NC-R | Y = −3.33 X + 0.72 | 0.59 | |
AOT30 vs. NC-S/NC-R | Y = −5.06 X + 0.72 | 0.62 | |
AOT40 vs. NC-S/NC-R | Y = −8.33 X + 0.72 | 0.65 | |
AOT60 vs. NC-S/NC-R | Y = −0.08 X + 0.73 | 0.68 | |
AOT80 vs. NC-S/NC-R | Y = −1.26 X + 0.71 | 0.65 |
Year | Regressors | Equation | R 2 |
---|---|---|---|
1999 | AOT40 vs. NC-S/NC-R (F.W.) | Y = −9.78 * 10−5X + 1.35 | 0.95 |
AOT40 vs. NC-S/NC-R (D.W.) | Y = – 9.78 * 10−5X + 1.32 | 0.95 | |
2000 | AOT40 vs. NC-S/NC-R (F.W.) | Y = −3.74 * 10−5X + 1.01 | 0.88 |
AOT40 vs. NC-S/NC-R (D.W.) | Y = −2.69 * 10−5X + 0.85 | 0.84 |
Extrapolation of responses of these sentinel plants to natural and semi-natural vegetation is difficult because of differences in genetic structure, community composition, and especially environmental adaptations.49 Native (detector) plants, growing in situ, are preferred for vegetation surveys. These established detector bioindicators visibly respond to O3 only when sufficient soil moisture and climatic conditions allow uptake of enough O3 for a sufficient time to allow inactivation of antioxidant defence mechanisms.50 As an alternative, brown knapweed (Centaurea jacea L.) has been suggested as a potential native bioindicator agent because this species shows characteristic symptoms of O3 injury in the form of bronze stipples or leaf discoloration after only a few days of field exposure to elevated O3.51 Symptoms were found to be identical to those observed in open-top chamber experiments.72 Similarly, Manning et al.50 proposed the closely related Centaurea nigra L. as a bioindicator species in central European forests. Due to its broad climatic adaptation, C. jacea is indigenous in many European countries. Recently, the above mentioned European project introduced a common protocol involving C. jacea plants as a model species to evaluate the relationship between ozone exposure and effects on the performance and injury symptoms of native plant species.42 As expected, a strong intraspecific variation in ozone response has been detected in this sensitive wild species:52 the extent of visible injury under identical environmental conditions varies among populations from different regions with different climates and pollution histories, and differences in plant ontogeny.
Searching for an alternative O3 bioindicator, another Solanacea has been tested, Lycopersicon pimpinellifolium, also known as the currant tomato. This is a wild plant (0.5–1 m height) native to western South America, that can be easily cultivated in any kind of drained soil, and only requires a sunny position.53 Though its sensitivity to O3 was described by Gentile et al.,54 no other reports have been published since then on the level and traits of this sensitivity.74 Besides confirming that L. pimpinellifolium is a highly O3-sensitive species, it is suggest that it can be positively evaluated as an effective bioindicator of the pollutant.
It is known that the Mediterranean vegetation may tolerate potentially harmful O3 levels.73 The prevailing weather conditions in the Mediterranean area (excess light, elevated temperature, reduced precipitation) reduce gsto at the time of the highest O3 levels, and promote sclerophylly, VOC emission and antioxidant activity. These avoidance and detoxification mechanisms may be of greater importance for Mediterranean vegetation than for that from central and northern Europe.
Quality assurance (QA) is still a relatively new topic in biomonitoring. QA is ‘an organized group of activities defining the way in which tasks are to be performed to ensure an expressed level of quality’. This means that all steps of the programme should be correctly addressed, from the design of the survey to data collection, processing and reporting. Four main activities are considered in a QA program: Quality Management (QM), whose major task is to ensure that activities are performed in a proper way; Quality Assurance (QA), whose task is to provide consistent methods with verified data quality; Quality Control (QC), whose major task is to ensure that data are appropriately collected and quality assurance is carried out; Quality Evaluation, which allows precision and accuracy of determinations to be evaluated, providing a basis to evaluate the comparability of data. A brilliant example of management of data quality in the environmental field is given by American and European forest health monitoring programs.56,57
QA procedures were adopted to ensure proper design and implementation of the study conducted in Florence.30 Results obtained before the beginning of the campaign were reassuring and, in terms of data quality limits, 90% of the scores of injured tobacco Bel-W3 leaves examined fell into the exact class or in the adjacent ones.
One of the advantages in biomonitoring is in measuring rapidly the magnitude of the injury. The impracticability of counting and measuring individual lesions on indicator plants motivated research workers to rely on rapid visual methods with the help of iconographic material and pathometric scales. The observer is a significant source of measurement error. Large scale campaigns, which involve several observers, are exposed to significant risks of estimation errors. Because disease scoring apparently follows the Weber–Fechner law, which states that the response of an organism to a stimulus is a linear function of the logarithm of the stimulus, graded readings are often converted to percentages on a logarithmic curve.58 Psychological studies provide evidence that the eye often grades inaccurately.59 Shape, orientation, shading, surrounding figures and personal traits enter into perception.
Visual assessments must be made quickly and do not require expensive equipment, chemical analysis or highly trained personnel, but their subjective nature creates concern. For ozone-induced injury, various attempts at improving the reading procedures have been reported.60 Mortensen described an integrated unit which illuminates the leaves from behind and allows the taking of standardized photos which are analyzed by a computer.60 Della Mea et al.61 presented a method of evaluation by computerized image analysis. However, the applicability of these techniques for large-scale surveys is doubtful due to operational limitations.
A specific study has been performed in Pisa.62 Fifty volunteers were selected to evaluate the accuracy of the visual method for assessing lesions induced by ozone on Bel-W3. Results indicate an easy estimation of the highest and lowest classes and a general difficulty in evaluating intermediate ones.
The visible foliar response of tobacco Bel-W3 can be used at least as a qualitative indicator of ambient ozone pollution.66,67 One of the reasons for the difficulty in relating ozone injury on adult plants to concentrations may be that prolonged exposures to ambient oxidants modify the response to subsequent exposures, when the same plants are used for several weeks; in the same way, the behaviour of plants immediately after being transported to the sites may be influenced by the radical change in environmental conditions. The adoption of methodologies based on short exposure times (i.e. 7 days, as in the case of the miniaturized kit) proved to be very useful. The effectiveness of the traditional biomonitoring method, based on the use of adult plants, was compared with the innovative miniaturized kit of seedlings: correlation studies among cotyledonar and foliar injury index and ozone levels monitored through physico-chemical measurements and results of analysis of variance emphasized the high degree of sensitivity of the innovative methodology.22
The mass media are very interested in spreading information on biomonitoring. Knowledge provided by indicator plants has incomparable value from the didactic and educational points of view: the vision of the severe macroscopic injury which ambient air causes to sensitive organisms may stimulate in the citizen a deeper involvement in environmental issues. Several pilot study projects have been launched with selected groups of teachers.
A standardization of methodologies, even at an international level (e.g. EU) is strongly felt to be needed, in order to get official acknowledgement of biomonitoring procedures. A first attempt of the Italian Environmental Agency (ANPA, Rome) to describe fundamental standard procedures of bioindication of ozone (from seed acquisition to plant raising, to personnel training and data analysis) has been successfully performed.68 International co-operation could allow the setting up of an extensive and inexpensive network for monitoring biological effects of air pollutants over large areas. As an ozone climatology for Europe and the Mediterranean basin is needed, a coordinated effort is desirable in order to integrate physico-chemical and biological data concerning the spatio-temporal distribution of this pollutant.
In Italy a fundamental reorganization of the national and regional health and environmental services has been launched: this could be a good opportunity for organizing large scale integrated surveys, where conventional monitoring facilities are coupled to biomonitoring units. A nation-wide monitoring network which gives complete coverage of the country should be organized, by implementing the present availability of analyzers and integrating it with biomonitors. A task-force should be put in charge of supervising all the aspects of the problem, including meteorological and biological ones, following the example first offered by the United Kingdom Photochemical Oxidants Review Group69 and then by the National Expert Group on Transboundary Air Pollution.70 It is important to continue discussion concerning the future use of tobacco, both as an ‘ozone monitor’ and as an indicator of ozone stress on other types of vegetation. The use of a biomonitor less sensitive than Bel-W3 tobacco could be envisaged, due to the very extreme response that this indicator systematically gives under the Italian pollution situation.
A glance to the future: selected perennial plants could be individualized to form permanent plots. The differential response to ozone is characterized in two poplar clones, indicating that the differences in phenomenological response are the effect of biochemical and physiological differences.71 In addition, biosensors and genetic engineering open up new application fields.
Data quality assessment should be a rule in biomonitoring activities. The quality of biological measurements, as with physico-chemical methods, depends on the reproducibility of the results. The degree of error intrinsic in the variability of all biological data is compensated for by the density of sampling points; this permits the construction of reliable pollution maps of large areas that could not have been obtained by instrumental recording due the scarcity of sampling points. This is particularly true in the case of the miniaturized kits, which allow an adequate sampling within the study site.
Biomonitoring, however, must overcome the present limitations, which include: a poor understanding of its validity, importance and versatility by the decision-makers, which may also be related to the poor ‘commercial’ inclination of the biomonitoring advocates; enthusiasm from the local environmental protection agencies, which—as a rule—are managed by chemists, who approach bioindication with scepticism; the lack of true standardization and formalization of the bioindication methodologies, as well as of a sound procedure for selection, instruction and evaluation of the field operators. The use of simple techniques, such as the miniaturized kit developed in Italy, should be encouraged. Basic research should be intensified, field courses and seminars organized: biomonitoring should be more ‘visible’!
Footnotes |
† Presented at the Fifth International Symposium on Modern Principles of Air Monitoring & Biomonitoring, June 12–16 2005, Norway. |
‡ The complete national dataset of ozone biomonitoring experience is available at http://www.biomonitoraggio.org. |
This journal is © The Royal Society of Chemistry 2006 |