Journal Archive

Black carbon (BC) is emitted from the combustion of fossil fuels and biomass (Ramanathan and Carmichael, 2008). BC produces a warming effect in the atmosphere by absorbing solar radiation and has a negative impact on human health (WHO, 2012; Forster et al., 2007). The major sources of BC are developing nations in the tropics and East Asia (Ramanathan and Carmichael, 2008). It is projected that BC emissions in Asia will continue to increase (Ohara et al., 2007). East Asian countries are facing the problem of transboundary air pollution, including BC and other aerosol particles, and its effects on vegetation (Izuta and Funada, 2010). Therefore, it is necessary to clarify the effects of BC particles on the growth and physiological functions of East Asian forest tree species.

Although there is little information on the effects of BC on the growth and physiological function of plants, it has been suggested that the BC deposited on the foliar surface at >400 mg C m^-2 projected leaf area increases leaf temperature and intercepts the irradiation light, which results in a reduction in net photosynthesis (Hirano et al., 1995, 1991). Although many researchers have reported deposition amounts of insoluble particulate matter and their elemental composition on foliar surfaces (e.g. Sæbø et al., 2012; Dzierzanowski et al., 2011; Takamatsu et al., 2001, 2000; Beckett et al., 2000; Turunen et al., 1997), there is no information on the specific amount of BC particles deposited on foliar surface. To clarify the effects of BC particles on growth and leaf or needle gas exchange rates of forest tree species, it is necessary to make field estimates for the amount of BC particles deposited on foliar surfaces (e.g. Matsuda et al., 2012). Furthermore, it is important to validate (using actual measurement) the estimated amount of BC deposited on the forest canopy. However, a large number of sample measurements are required for validation because of the huge and seasonal variation in the amount of BC deposition (Matsuda et al., 2012). Therefore, it is necessary to establish a simple, fast and cost-effective technique for measuring amount of BC deposited on the foliar surface of forest trees.

Actual measurement of atmospheric BC has been performed by thermal elemental carbon (EC) analysis (e.g. Schmid et al., 2001). It is not, however, realistic to carry out the large number of required measurements because the thermal optical method takes too long. On the other hand, an optical method using an integrating sphere was established and has been used for the measurement of BC (Hitzenberger et al., 2006, 1996). This method allows the measurement of spectral absorbance in a short time, and thus, allows the measurement of a large number of samples. It has been reported that light-absorbing compounds, such as brown carbon, affects the accuracy of the optical method for measuring BC amount (Reisinger et al., 2008). However, accuracy could be improved by using an appropriate wavelength based on the absorbance spectrum and by correction using a calibration curve with actual measurements using the thermal optical method. In the present study, therefore, we investigated the optical method for measuring the amount of BC extracted from the foliar surface of forest tree species and collected on a quartz fiber filter to analyze a large number of samples within a short time.

The relationship between the amount of BC in chloroform applied to a quartz fiber filter and the amount of BC collected on the filter, determined as that of EC, is shown in Fig. 1. There was significant correlation between the amount of BC applied to the filter and the amount of EC on the filter. The slope of the regression line was 0.917. This result indicates that the collection efficiency of BC particles suspended in chloroform was greater than 90%.

Fig. 1. 
The relationship between the amount of BC in chloroform applied to quartz fiber filter and the amount of BC collected on the filter determined as that of EC. Solid and dashed lines indicate a 1 : 1 line and the regression line of the relationship, respectively. The regression line was obtained from the linear regression analysis with the intercept forced through zero. The coefficient of determination (R²) and slope (±standard error) of the regression line were 0.991 (p<0.001) and 0.917 (±0.018), respectively.

Fig. 2 shows typical absorbance spectrums of the quartz fiber filter used for the collection of BC particles from the foliar surfaces of F. crenata, C. sieboldii, L. kaempferi and C. japonica seedlings. In all the tree species, the absorbance of the filter increased when the wavelength at which the absorbance was measured decreased. At a wavelength below 450 nm, the increase in the absorbance was greater than that above 450 nm. This tendency was observed in absorbance spectrums of all filters used for filtration of chloroform washings but was not observed for that of deionized water washings. More than 95% of the ultraviolet radiation (UV, wavelength below 400 nm) is absorbed by cuticular, suberized outer layers of the epidermis and pigments, such as flavonoids, in plant leaf (Larcher, 2003). Because the chloroform used for the extraction of BC particles dissolves the epicuticular wax of the leaves and needles (Cape et al., 1989), the absorption at a wavelength below 450 nm might be due to the epicuticular wax. In addition, in the seedlings of C. sieboldii, L. kaempferi and C. japonica, the increase in the absorbance below 450 nm tended to be greater than that in the F. crenata seedlings (Fig. 2). The species difference in the wavelength-dependent absorbance was observed in all filters used for filtration of chloroform washings. It has been suggested that the amount and composition of epicuticular wax changes the UV absorptivity of wax (Sase et al., 1998). Therefore, the species difference in the degree of the increase in the absorbance at a wavelength below 450 nm might be due to a difference in the amount and/or composition of the epicuticular wax among the species. It is important to identify the light-absorbing compounds and their optical absorption property for further application of the present method to field measurement practices. However, the aim of the present study can be achieved by selecting an appropriate wavelength to avoid the influence of absorption by light-absorbing compounds on the accuracy of the present optical method. Therefore, absorbance at a wavelength above 450 nm was used for the quantification of the amount of BC collected on the filter.

Fig. 2. 
Typical absorbance spectrums of quartz fiber filter used for the collection of BC particles extracted from the foliar surface of F. crenata (○), C. sieboldii (△), L. kaempferi (◇) and C. japonica (▽) seedlings.

Fig. 3 shows the relationship between wavelengths, at which the absorbance of quartz fiber filters was measured, and (a) the coefficient of determination (R²) and (b) the intercept of the regression line between the amount of EC on the filter (M_EC, μg C cm^-2 filter area) and the absorbance of the filter. The linear regression analyses were performed using the collective data of four tree species and two sample types (water and chloroform extract) due to the small number of data per species and sample type. The R² value increased drastically up to a wavelength of 450 nm and subsequently increased gradually. This tendency might be because of the absorption by biological substances such as epicuticular wax (Fig. 3). This result suggests that biological substances soluble in chloroform could hamper the relationship between M_EC and absorbance of the filter. Therefore, for more accurate estimation, it is necessary to confirm the absorption spectrum of the filter for each tree species to clarify whether there is an absorption maximum originating from these biological substances.

Fig. 3. 
The relationship between the wavelength, at which the absorbance of quartz fiber filters was measured, and (a) coefficient of determination (R²) and (b) intercept of the regression line between the amounts of EC on the quartz fiber filter used for the collection of BC particles extracted from the foliar surface and the absorbance of the filter.

All the R² values were statistically significant (Fig. 3a). Relatively high R² values (above 0.9) were observed at wavelengths above 435 nm, and the highest value was observed at a wavelength of 600 nm (Fig. 3a). While increasing the wavelength, the intercept of the regression line drastically decreased up to a wavelength of 450 nm and subsequently decreased gradually (Fig. 3b). Because the intercepts were not significantly different from 0 above 425 nm (data not shown), the absorbance of the filter at wavelength above 425 nm was statistically applicable to the estimation of the amount of BC particles collected on the filter. However, for more accurate estimation, it is desirable to use the absorbance at a wavelength above 450 nm because of absorption by epicuticular wax (Fig. 2). Furthermore, the intercept of the regression line between M_EC and the absorbance of the filter at 580 nm was closest to 0 (Fig. 3b). Ideally, the intercept of the regression line should be 0. Therefore, we assumed that the absorbance at 580 nm was suitable for estimating the amount of BC particles collected on the filter.

The relationship between M_EC and the absorbance at 580 nm (A₅₈₀) is shown in Fig. 4; a significant linear relationship (R²=0.917, p<0.001) is obtained with a coefficient of variation for the slope of 6% as A₅₈₀=(0.052±0.003) M_EC-(0.000±0.017). From the standard error (σ=0.017) of the interception of the relation, the detection limit (3σ) and limit of quantification (10 σ) of M_EC were calculated as about 1.0 and 3.3 μg C cm^-2, respectively. The 95% confidential intervals of the regression line were about ±0.4 μg C cm^-2 around 4 μg C cm^-2, ±0.5 μg C cm^-2 around 7 μg C cm^-2 and ±1.0 μg C cm^-2 around 10 μg C cm^-2. These results indicate that the method can predict M_EC with an error of about 10% for moderately loaded filter samples (0.2<A₅₈₀<0.5), and filter loading of 4 to 10 μg C cm^-2 is the useful range of the prediction. The amount of light-absorbing compounds will vary depending on the amount of foliar samples, composition of biological substances and amount of contaminant deposited on the foliar surface. Therefore, it is necessary to confirm the appropriate wavelength and an applicable range of the relationship for each sample type, as determined by tree species and harvest region. This method, however, may serve as a simple, fast and costeffective technique to measure amount of BC particles on foliar surface.

Fig. 4. 
The relationship between the amount of EC on the quartz fiber filter (M_EC, μg C cm^-2 filter area) used for the collection of BC particles extracted from the foliar surface and the absorbance at 580 nm of the filter. The relationship was obtained by linear regression analysis using the collective data of the four tree species. The dashed lines indicate the 95% confidential intervals of the regression line. The coefficient of determination (R²) was 0.917 (p<0.001) and slope and intercept (±standard error) of the regression line were 0.052 (± 0.003) and -0.000 (±0.017), respectively. ○: F. crenata, △: C. sieboldii, ◇: L. kaempferi, ▽: C. japonica.

The results of the present study demonstrated that the amount of BC particles extracted from foliar surfaces of forest tree species and collected on a quartz fiber filter can be estimated from the absorbance of the filter. Part of the results were obtained using artificial BC particles, therefore, it is unclear whether the collection efficiency and calibration curve reported are directly applicable to the leaves or needles of forest trees in the field. However, this optical method can serve as a simple, fast and cost-effective technique for field measurements of amount of BC deposited on the foliar surface of forest trees.