LEAF SPECTRAL BEHAVIOR AND CHLOROPHYLL CONTENT OF MIMOSA HOSTILIS CANOPY IN A SEMIARID ENVIRONMENT

Insurance and efficiency to monitoring and the use of interacting methods for an accurate diagnosis of the vegetation, mainly those from tropical regions, are essentials to contribute to future development management actions. Remote sensing, biological traits associating structural and biochemical data will permit evaluate the behavioral picture of a dominant specie in vegetation that cover great areas. This study aimed to identify the leaf spectral signature features and structural traits related to dominant specie and photosynthesis implications, relating it with environmental stress consequences and discussing critically monitoring using remote sensing methods. Multispectral analysis and hyperspectral spectroradiometer “FieldSpec Handheld” by Analytical Spectral Devices (ASD) from georeferenced points were made. Perceptible differences were detected for mean values of chlorophyll a and b from field data. The use of Hyperion hyperspectral image showed potential to infer the condition of vegetation regeneration analyzed and different intensities of chlorophyll.


Introduction
Remote The chlorophyll quantification, mainly those related to the vegetation with few species dominant assumes greater relevance when the intention is to evaluate its biological condition.Different crop species were evaluated through chlorophyll quantification and hyperspectral indices of the canopy, in an area with different sizes (Ferri, Formaggio & Schiavinato, 2004;Jarocinska & Zagajewski, 2009).
The literature shows different approaches to quantify the chlorophyll and hyperspectral indices of vegetation to inform, to estimate and to produce models that can inform about the health of plants photosynthetic performance (Haboudane et al., 2004;Ferri, Formaggio & Schiavinato, 2004;Malenovský at al., 2006;Wu et al., 2008).
Figure 1.Map of the study area highlighting the studied area.

Multispectral data
Basin for the multispectral analysis was a subset of Landsat 5 satellite image of January 13 nd , 2009, the satellite was launched in 1970 and is still collecting reflectance data 180km above ground.The spatial resolution of 30m in 1, 2, 3, 4, 5 and 7 bands and 120m in thermal band.Landsat 5 collects multispectral data in seven spectral bands.Table 1 shows the band numbers and width of the satellite.

Hyperspectral data
The hyperspectral spectroradiometer Additionally, a low cost GPS solution coupled to a CE computer was used to locate the sampling spots (Laudien et al., 2004).
A white reference plate under the sheet was used to simulate the effect of background.

GPS data
GPS-data was collected at selected fields in all points of the study area to validate the multi and hyperspectral analysis results.
Multispectral or hyperspectral vegetation indices can be calculated by using the red and near-infrared portions of the spectra as well (1) Where, R750 = reflectance at 750nm (%) and R700 = reflectance at 700nm (%) (2) Where, RB4 = reflectance of the band 4 and RB3 = reflectance of the band 3.
In the last step, the results of the VI calculations were classified into nine vitality classes by using the "Quantile Classification Method" of ArcGIS TM Spatial Analyst.To  (January 23, 2010).

Spectral response of the Landsat TM 5
The Figure 3

Spectral response using Hyperion
The

Conclusions
The Landsat contributes to obtain optimal estimations to the vegetation index compared with information from high spectral resolution of the Hyperion and FieldSpec.
Field data permitted find differences for mean values of chlorophyll a when compared with chlorophyll b.

Authors
sensing data is widely used for canopy monitoring in different vegetation types.Though it is possible to analyze biochemical characterization of plants in different wavelengths of the electromagnetic spectrum, apparently few is known about spectral traits and chlorophyll content into leaves in tropical plants.It can be also used to modeling and to simulate biological and physical processes.

"
FieldSpec Handheld" by Analytical Spectral Devices (ASD) was used to collect field data of an artificial inoculation trial.This very high spectral resolution device was applied to evaluate biotic and abiotic growth-anomalies of sugar beets in dependence of their spectral reflectance.The used ASD handheld spectroradiometer collects spectral data in a wavelength range of 325nm to 1075nm with an interval of 1.6nm.This results numerous individual bands which represent a nearly contentious reflectance curve.The viewing angle of the FieldSpec is 25 degrees.

(
Laudien et al., 2004).The index values are significantly correlated to the vitality of the detected plant.In this study, the hyperspectral vegetation index (HVI), following Gitelson et al. (1996), was modified according to Laudien et al. (2004) (see Equation 1-3).Equation 2 presents the one modified spectral vegetation indices (VI).The spatial analysis, both calculation and classification, were accomplished by using the GIS software Erdas Image TM 9.3 by Leica Geosystems ® and ArcGIS TM 9.3 by ESRI ® .
validate and compare the two classified outputs, the GPS polygons of the campaign were to the ArcGIS TM 9.3 project as an overlay layer, proposed by Laudien et al. (2004).Aiming to relate hyperspectral spectroradiometer features with internal structure of the leaves of M. hostilis we measured the traits: thickness of upper and lower cuticle and epidermal layers, mesophyll, palisade and spongy parenchyma, intercellular air spaces, vascular tissues and substomatal chamber to evaluate the internal backscattering of light inside the leaf.These measures will permit discuss the degree participation of each one with the reflectance and transmittance processes.the medium spectral behavior in the visible of M. hostilis Benth.to the point from the coordinates x=738057 and y=9245952 meters.Note that the spectral reflectance in the green band was approximately 27% and in the red band around 20%.The HVI average was 2.002.At this point, the average value for chlorophyll a was 21.7, and to b was 3.5, on January 23, 2010.

Figure 2 .
Figure 2. Spectral response of Mimosa hostilis measured with FieldSpec in laboratory conditions.
shows the spectral behavior of M. hostilis in a semiarid region of Northeast Brazil using the satellite Landsat5.In general, the different samples showed a similar pattern of spectral behavior.However, there are perceptible differences for the intensities from reflected energy.The red line coincides with the green line; it showed the same behavior in all spectral bands of the electromagnetic spectrum.Photographs 1 and 2 of the landscape correspond to the points 1 and 2, for red and green lines in the Figure 3.The point 3 corresponds to the black line, which exhibited different intensities of energy reflected from other points in the visible region, and it was approached the energy reflected from the points 1 and 2 to the near infrared and middle regions.It is possible note in the point 3, a lower intensity of reflected energy in the visible when compared to the points 1, 2 and 5.The point 4 showed lower intensities of reflected energy.Photographs 4, 5 and 6 show the landscape related to the points 3, 4 and 5, corresponding to the lines black, yellow and purple, respectively.The points 1, 2 and 3 showed similar features relating to cell structural traits in individuals from different species that have similar development stadium.These points correspond to the vegetation regeneration, indicating that the analysis allows us to infer the condition of vegetation regeneration analyzed and different intensities of chlorophyll due to higher density of green coverage in the point 3. Through field sampling was possible verify differences in mean chlorophyll a of 8.3, 5.7 and 18.8 and chlorophyll b of 2.2, 3.0 and 4.7 in the points 1, 2 and 3, respectively.Section 4 showed the least amount of reflected energy (yellow line) and the highest average amounts of chlorophyll a. Section 5 showed the highest energy level reflected in all regions of the electromagnetic spectrum.The VI Landsat 5 average was 1.76, with a minimum of 1.38 and a maximum of 2.17.Considering the average value and comparing the value obtained with the FieldSpec, we found that it represents 88% from the value obtained with the spectroradiometer.It means that even with all environmental interference and variations of the characteristics in the sites studied, the values of vegetation indices obtained from satellite images represent very well the actual percentage of vegetation in the area.
using of FieldSpec under field conditions associated with chlorophyll content data of M. hostilis individuals in a semiarid environment in a Hyperion hyperspectral image showed interesting results.It was emphasized when we collected data under different environmental conditions (Figure 4).The average of VI Hyperion was 1.35, representing 68% from the vegetation index obtained through the FieldSpec and 77% from that obtained with the Landsat TM.It was noted that, even at low spectral resolution, Landsat provides optimal estimates for the vegetation indices in relation to the M. hostilis.

Figure 4 .
Figure 4. Photographs showing the experimental areas with dominance of Mimosa hostilis, and the five points used to collect data using FieldSpec and GPS.(November 15, 2010).