Vitis vinifera SPECTRAL RESPONSE TO THE INCREASE OF CO 2

Hyperspectral remote sensing (HRS) is a useful method to monitor spectral changes in vegetation. HRS contains significant spectral information for detecting plant stress. The specific aims were: (1) to assess the changes in Vitis vinifera plant chlorophyll content due to the leakage of CO2 into the plant-air environment, and (2) to analyze an vegetation index derived from the first derivative reflectance values for use in detecting Vitis vinifera plant stress due to elevated concentrations of air CO2. Spectral reflectance was measured between 336 and 1045 nm with a spectral resolution of 1 nm, covering visible and near-infrared portions of the electromagnetic spectrum. The amount of chlorophyll decreased about 50% in the open top chamber modified (OTC modified) + CO2 injection when compared to natural condition. The difference in chlorophyll between OTC modified + no CO2 injection and natural condition was 24%. The concentration of chlorophyll a and b decreased and concentration of carotenoids increases of Vitis vinifera in initial stage of growth, with increase in CO2 to 550 ppm. In the end, the remote sensing hyperspectral presents itself as a great tool to assist in studies of global climate change and its impacts on the biomes of the world.


Introduction
Much has been discussed about global climate change, greenhouse gases, global warming, and its possible effects in Brazilian agriculture.A recent study coordinated by Pinto & Assad (2008), indicates that the increase in temperature can cause, in general, a decrease in regions with lower climate risk for the grain cultivation and significant economic losses (Coltri et al. 2009).
One of the potential options to mitigate the contribution of fossil fuel emissions to global warming problem is to CO 2 capture and store (Lakkaraju et al., 2010;IPCC, 2005).
However, associated with the knowledge on CO 2 storage the extent of the CO 2 leakage is one of key questions related to integrity of storage (Lakkaraju et al., 2010;Hepple and Benson, 2005).

Solar radiation interacts with
vegetation through the leaves, first with leaf surface and then with mesophyll; and this leads to integration between cells, tissues, pigments, secondary compounds of cellular metabolism, energy in the light photon, and incidence angle of rays (Pimentel, 2010).
While it has long been known that photosynthetic pigments control the visible reflectance properties of leaves, only recently, techniques have been developed to estimate the concentrations of individual pigments within leaves, using high spectral resolution (hyperspectral) reflectance measurements (Blackburn, 1998;Chappelle et al., 1992;Peñuelas et al., 1995).
The distribution of pigments within a leaf is associated with format and size of cells and number and size of intracellular spaces.
Palisade cells have a much greater number of chloroplasts in relation to the amount of these structures found in the foam, making the palisade a tissue with higher absorbance values in the mesophyll.Consequently, the larger the area occupied by cells of palisade in mesophyll, it is expected that a higher absorbance value is obtained in this leaf (Pimentel, 2010).
Hyperspectral remote sensing (HRS) is a useful method to monitor spectral changes in vegetation.HRS contains significant spectral information for detecting plant stress (Lakkaraju et al., 2010;Carter, 1994;Carter, 1998).Spectrometry imaging has been used by De Jong (1998) to detect CO 2 seepage indirectly, using vegetation spectral reflectance.
Photosynthetic pigments such as chlorophyll a and chlorophyll b (which absorb photons that energize the reactions of photosynthesis) and carotenoids (which protect the photosynthetic reaction centers from the light excess) dictate the photosynthetic potential of a leaf and relate strongly to the physiological status of plants (Blackburn, 1998).
Wine production in Brazil had considerable increase from the nineteenth century with the arrival of Italian immigrants in the Rio Grande do Sul state.The Brazilian viticulture has evolutes in a extraordinary way in the two last decades, with a considerable increase in area planted with Vitis vinifera, giving to Brazil an international visibility as a producer of fine wines.

Location and Climate
The experiment was carried out at the Embrapa Tropical Semiarid Experimental Field (09º09 'S latitude and 40º22' W longitude), located in Petrolina municipality (Figure 1) in 15 th November 2010.
The results were presented by averaged spectral curve for each treatment.Three treatments (with CO 2 and without CO 2 in the OTC modified and control (attestant) were used.For each treatment were collected nine measurements.The PSSRa index was 8.51 with OTC modified + CO 2 and 16.52 with OTC modified + no CO 2 injection Table 1.There was a decrease in chlorophyll a with CO 2 injection.

Results and discussion
The PSSRb was 6.10 with OTC modified + CO 2 and 11.89 with OTC modified + no CO 2 .
There was a decrease of chlorophyll b with The SIPI was 1.05 on OTC modified + CO 2 injection and 1.02 on OTC modified + no CO 2 injection, (Table 1).In the OTC modified + CO 2 injection there was an increase in amount of SIPI, i.e. an increase in ratio between chlorophyll a and carotenoids.There   The values of PRI were negative for natural condition and OTC modified + CO 2 injection, and ranged between -0.07 and -0.004 (Table 1).The vegetation was performing photosynthesis with efficiency in the use of active radiation at the moment of collection.
Thus, due to realization of photosynthesis, it is clear that the Vitis vinifera was using CO 2 with high efficiency.When collecting data on OTC modified + no CO 2 injection the vegetation was not performing photosynthesis, but  1981, Gamon et al. 1995;;Blackborn, 1998).found that chlorophyll content in gas exposed to high concentrations of soil natural gas was decreased by 30%.In addition, Ketel (1996) found that the chlorophyll content of leaves, but treated with 25% lethat dose of glyphosate was equal to that of control leaves, but that when lower doses of the herbicide were used, the chlorophyll content was greater that of control leaves.In Lakkaraju, (2010) as CO 2 seeps to the ground surface, it could deplete oxygen in the soil atmosphere and could cause stress in the local vegetation.The primary cause of the stress, in response to natural gas leaks, is believed to be displacement of oxygen from the soil atmosphere, which thereby inhibits root respiration that provides energy for root growth and uptake of nutrients from the soil (Hoeks, 1972a;Gilman et al., 1982;Arthur et al., 1985).Displacement of soil oxygen has negative effects on plant growth which are expressed as reduced root, shoot growth and dry weight (Drew, 1991).For instance, Noomen and Skidmore (2009) (1994), this increase in average global temperature will cause in the higher latitudes of the northern hemisphere an increase of up to 12 o C in mean annual temperature, which will likely determine collapse of vegetation such as the Tundra and Forest.

Conclusions
The spectral indices used in this study proved to be excellent for estimating concentration of photosynthetic pigments (such as chlorophyll a and b and carotenoids) and they were sensitive to increased CO 2 and its impact on response of Vitis vinifera.Gilman, E., Leone, I., Flower, F., 1982.Influence of soil gas contamination on tree root health.Plant and Soil 65, 3-10.
vine appears as the third most important crop in terms of planted area, with an estimated area of 8,000 hectares in 2002, surpassed only by the areas planted with mango and coconut (Silva & Correa, 2004).It is worth mentioning the scarcity of statistics on horticulture in the region.The vine crop is of special economic and social importance, with involves a large annual volume of businesses directed to internal and external markets and stands out among the irrigated crops in the region such as those that present the highest coefficient generating direct and indirect jobs (Silva & Correa, 2004).Remote sensing data in situ are often colleted by researchers simply to obtain information about the material spectral characteristics of terrestrial surface.Additional knowledge on the materials can be obtained submitting to various treatments (e.g.applying different amounts of CO 2 in the air to evaluate the response of Vitis vinifera) and determining whether treatments result in different patterns of spectral reflectance.The several treatments and their characteristics can be monitored over time to get additional information.In situ reflectance data from nearly pure materials of Earth's surface can be used to locate endmembers for use during the analysis of hyperspectral or multispectral data (Jensen, associated with other remote sensing data to identify their stresses, productivity, and other hydrological variables of vegetation.Thus, many studies of the remote sensing are focusing on monitoring of what happens with the photosynthetically active radiation (PAR) when it interacts with the individual leaves or the vegetation canopy.The use of imaging and non imaging spectroradiometer is particularly useful for measuring characteristics of absorption and reflectance of photosynthetically active radiation (Jensen, 2009).The aim of this study was to use hyperspectral remote sensing techniques to investigate the spectral responses of Vitis vinifera to increased CO 2 concentrations in atmosphere air, where CO 2 was injected into air of the OTC modified, which show 2 m of diameter and 1,2 m height.The specific aims were: (1) to assess the changes in Vitis vinifera plant chlorophyll content due to the leakage of CO 2 into the plant-air environment and (2) to analyze an vegetation index derived from the first derivative reflectance values for use in detecting Vitis vinifera plant stress due to elevated concentrations of air CO 2 .

Figure 1 -
Figure 1 -Spatial location of the site.

Figure 2 -
Figure 2 -Climatology of the air temperature and precipitation for Petrolina municipality, Pernambuco State, Brazil.
structures called OTC modified (Open Top Chamber modified).The OTC modified are circular structures, framed with aluminum rods and side of PVC film (polyvinyl chloride), with 2 m diameter and 1.2 m height, Figure 3.The CO 2 concentration of was measured by infrared gas analyzers (IRGA, infrared gas analyzer), which provided information to a controller that regulated opening of valves for injection of CO 2 in the OTC modified.The CO 2 release was at a constant rate.

Figure 3 -
Figure 3 -Open Top Chamber modified (OTC modified) installed on the experimental site, 1998) were used to estimate the change in chlorophyll content of the plants and these indices were used to analyze the impacts of CO 2 increase in air on the photosynthetic pigment amount of plants.In the visible spectral region, the high absorption of radiation energy is due to leaf pigments; primarily the chlorophylls and carotenoids (Knipling, 1970) and therefore it would be possible to track changes in chlorophyll content by calculating vegetation indices in the visible spectrum.Since SIPI compares carotenoids with chlorophyll a, Chl NDI is an indicator of total chlorophyll content and PSSRa and PSSRb are the indicators of chlorophyll a and chlorophyll b, these indices were chosen to estimate changes in the concentrations of carotenoids, total chlorophyll, chlorophyll a and chlorophyll b.

)
respectively.Derivative spectra were calculated by differentiating the spectral reflectance with respect to wavelength.The wavelengths at which the first derivative spectra reach maximum and minimum values were used to derive a vegetation index to quantify the difference between CO 2 stressed and control vegetation.In the first derivative spectra, for CO 2 -stressed vegetation, the minimum (negative) was found to locate between 575 nm and 580 nm and maximum (positive) was between 720 nm and 723 nm.The wavelengths at these two features (minimum and maximum) were selected manually and used to obtain the Normalized difference First Derivative Index (NFDI).The averages of the first derivative values of spectral reflectance between 575nm and 580nm were then taken.The averages of the first derivative values of spectral reflectance between 720 nm and 723 nm were also calculated.NFDI was calculated as follows, based on these average values: where dR 575-580 is the average of the absolute first derivative values between 575 nm and 580 nm, and dR 720-723 is the averaged first derivative values between 720 nm and 723 nm.The percentage difference between stressed and control vegetation was calculated to assess the percentage impact of CO 2 injection in plant.A Incoterm digital laser thermometer was used to measure the temperature of leaf surface.

Figure 4
Figure 4 presents the daily trend of air temperature and relative air humidity in minute-scale measured in 15 th November 2010 at OTC modified plots in Petrolina municipality, Pernambuco State, Brazil.It can be observed that the for that day, the minimal air temperature was 22.78 o C at 08h35min (Greenwich) and the maximal one was 34.64 o C at 15h03min (Greenwich).

Figure 4 -
Figure 4 -Air temperature and relative air humidity in minute-scale measured in 15 th November 2010 at the OTC modified plots in Petrolina municipality, Pernambuco State, Brazil.

Figure 5
Figure 5 presents the daily trend of energy flux density of global solar radiation (Sr, W m -2 ) and photossynthetically active radiation (PAR, W m -2 ) measured in minutescale for 15 th November 2010 at the OTC modified plots.According to the Sr data, with the values rising up to 1196 W m -2 , this studied day was characterized by a great occurrence of clouds.Field measurements of the spectral data were done at a clear time of the day.PAR behavior is similar to solar global radiation, and by considering sunny day, the ration between photossynthetically active radiation and solar global radiation is around 50%.

Figure 5 -
Figure 5 -Energy flux density of global solar radiation (Sr, W m -2 ) and photossynthetically active radiation (PAR, W m -2 ) measured in minute-scale for 15 th November 2010 at the OTC modified plots in Petrolina municipality, Pernambuco State, Brazil.

Figure 6 -
Figure 6 -Grape Leaf surface temperature and air temperature at the spectral measurements time in vines growing in natural condition (T1), in OTC modified + no CO2 injection (T2) and, in OTC modified + CO2 injection (T3), in Petrolina municipality, Pernambuco State, Brazil.

Figure 7
Figure 7 shows the spectral response in visible region of Italy Vitis vinifera variety in three treatments: OTC modified + CO 2 injection, OTC modified + no CO 2 injection and natural condition.

Figure 7 -
Figure 7 -Spectral response in visible region of cv.Italy Vitis vinifera.
CO 2 injection.Blackburn (1998) applied a correlation between a data set of chlorophyll a and b with the indices of PSSRa and PSSRb, and found a strong correlation with a coefficient of determination from 0.93 to 0.94 for chlorophyll a and chlorophyll b, respectively.According to the author, PSSR index is used to minimize effects of interaction of radiation with the leaf surface and internal structure in the mesophyll, (Blackburn, 1998).
photochemical conditions and represents correlation between the epoxidation state of xanthophylls cycle pigments and photosynthetic radiation use efficiency.PRI for OTC modified + CO 2 injection was negative, (Table 1).The PRI index nearest of zero was the one of natural condition.The
photosynthesis.It is suggested that CO 2 interfere in the efficiency of plant photosynthesis.Indices SR705, mSR705 and Chl NDI which are indices to estimate the chlorophyll content showed the same trends, i.e., the OTC modified + CO 2 injection values decreased, (Table 1), confirming the results of the indexes PSSRa and PSSRb.In the analysis of Chl NDI there was a decrease in the rate of normalized difference chlorophyll with CO 2 injection.The Chl NDI is used to measure the fraction of Photosynthetically active radiation and coverage of green vegetation when used for large geographical scales (Kumar & Monteith,

Figure 9
Figure 9 shows the contents of chlorophyll a and b Vitis vinifera of the three treatments on November 15 th , 2010.Note that

Figure 9 -
Figure 9 -Change in chlorophyll content in three treatments and measured on November 15, 2010.
found that increasing soil CO 2 concentrations decreased plant height, leaf chlorophyll content and dry weight of maize plants.Boru et al. (2003) reported that a 50% CO 2 concentration at the root zone has shown either death of soybean plants or severe symptoms of chlorosis, necrosis and root death.Compromising to the health of plants due to stress is often associated with an increase in spectral reflectance in the visible region and a decrease in the near-infrared (NIR) region.This results in a shift of the slope between red and NIR, called "red-edge position", towards shorter wavelengths.Since it is known that the changes in spectral reflectance are often a stress induced response of plants (Macek et al., 2005; De Jong, 1998), these changes in spectral reflectance could be used as a proxy indicator of CO 2 leakage.First derivative analysis was applied to determine how red edge position was affected by air increase of CO 2 .Figure 10 shows the first derivate of reflectance for the Vitis vinifera with and OTC modified + no CO 2 injection.The injection of CO 2 caused stresse in plant.Note difference on reflectance between the response of Vitis vinifera with CO 2 and without CO 2 , especially in Red Edge region (region by 700nm).The indice Red Edge, (Table 1), shows lowest value in plant with CO 2 injection.In OTC modified + CO 2 injection occurred a increase relative compared with the natural condition in first derivative between 520 and 580 nm and a relative decrease in derivative between 630 and 660 nm.Similar result can be in Smith et al., (2005).Smith et al., (2004) examine the response of vegetation to the increase of methane and obtained a decrease of reflectance in the infrared.

Figure 11
Figure 11 shows the NFDI of OTC modified + CO 2 injection and natural condition.Note that the NFDI was less at and had a 9% difference.The NFDI was sensitive to stress of vegetation.According to the results, it is suggested that there is a negative correlation between increased CO 2 and NFDI.This information is of paramount importance for monitoring and forecasting of the impacts of increasing CO 2 in the air over the Vitis vinifera.Lakkaraju et al. (2010) found a high negative correlation (r =-0.93) between

Figure 11 -
Figure 11 -NFDI of the Vitis vinifera with CO 2 in air and natural condition.

Table 1 -
Spectral indices value computed in this study.