Laboratory observation of fluorescence spectra

407 Figure 5 shows the spatial distribution of SIF. In this image, the right side shows the reflectance of the WB, while the left side the reflectance from the soybean canopy. The camera image was recorded by applying the narrow bandpass F760 filter. We assume the linearity between the pixel digital numbers DNs and radiation intensity, and the same scaling factor obtained in the case of Fig. 4a is applied to this case as well. Thus, the difference between the canopy DNs and scaled reference DNs gives the SIF intensity distribution as indicated in Fig. 5b. The temporal change of the SIF intensity from both the spectral and camera measurements was compared with that of photosynthetically active radiation PAR recorded using a PAR meter BMS 3415FXSE Figure not shown. Good agreements were seen among these three independent datasets, suggesting the direct relation between the PAR and fluorescence intensities during the observation time. Figure 6a shows the reflectance image taken with the CCD camera attached with an interference filter having the central wavelength 550 nm and width 10 nm. The stand-off distance was 20 m, and the acquisition time duration was 30 ms. In this visible image, it is seen that generally strong reflection of solar radiation is observed for leaves that meet the specular reflection condition. Figure 6b, on the other hand, shows the SIF image taken by applying the F760 filter with the acquisition time of 50 ms. The color scale is assigned for the data expressed in units of count ms -1 pixel -1 . In the experimental field of Kyoto University, different varieties of soybean are planted column by column. Figure 7b, the red squares indicate the position of columns. As seen from Fig. 7, good agreement was seen for the SIF intensities determined from the spectral measurement and image analysis. In conclusion, we have described the SIF measurement from vegetation canopy under both laboratory and field conditions. The use of LED light source coupled with appropriate filter setup made it possible to separate the weak fluorescence intensity from much larger near- infrared reflectance of vegetation. In the field measurement, the scaling of the whiteboard reference worked well for deriving the fluorescence signals even under daylight conditions. Acknowledgement: This work was financially supported by the Grant-in-aid from MEXT and CEReS Joint Research. References 1 L. Guanter et al.: Geophys. Res. Lett., 34, L08401, doi:10.10292007GL029289 2007. 2 C. Frankenberg et al.: Geophys. Res. Lett., 38, L03801, doi:10.10292010GL 045896 2011. 3 P.J. Zarco-Tejada et al.: Remote Sensing of Environment 117, pp. 322-337 2012. 4 F. Daumard et al.: IEEE Trans. Geosci. Remote Sens., 48, pp. 3358-3368 2010. 5 K. Masuda Kuriyama et al.:WEP.N92, IGARSS Quebec 2014. 6 K. Kuriyama et al.: RSSJ 59th Autumn Meeting B-4 2015 a b Fig. 6 a Reflection image observed at 550 nm and b fluorescence image observed at 760 nm. Fig. 7 Difference in fluorescence intensity observed for different varieties of soybean 14:10-14:30 on September 2, 2016 Joint Scientific Symposium IJJSS 2016 Chiba, 20‐24 November 2016 408 THE USE OF HYPERSPECTRAL DATA TO ANALYZE CLIMATE CHANGE ACCORDING TO CARBON STOCKS AND SOUTHEAST SULAWESI BIODIVERSITY Sawaludin, Derick Christopher Ambo Masse, Muhammad Apdal, Hasmina Tari Mokui, Surya Kurniawan and Waode Nurhaidar, La Ode Restele Universitas Halu Oleo, Jl. H.E.A Mokodompit, Kendari 93232, Indonesia Abstract ndonesian region is lately experiencing extreme weather resulting in deviant climate change. t is occurred due to the massive amount of carbon dioxide produced on this planet. As the consequence, the ozone layer protecting the earth from solar radiation becomes vulnerable and damaged. The existence of the green plants classified as forest biodiversity is expected to reduce the large amount of carbon dioxide gas by transforming it into oxygen through photosynthesis. Forest biodiversity in Southeast Sulawesi is categorized into tropical forest, savannah and mangrove forest. Each type of forest has different amount of carbon stock. Nowadays, , information in relation to carbon stocks and forest biodiversity are needed for the analysis of climate change. The main objective of this research is to analyze climate change according to carbon stocks and forest biodiversity in Southeast Sulawesi by applying hyper spectral data. n order to distinguish the forest types and to comprehend its carbon potentials, unsupervised classification method is applied and correlated with the biomass carbon. Furthermore, it is analyzed using NDV formula to determine the extent and density of vegetation and carbon potential existing in an area. Further analysis is conducted by utilizing weather parameter of air pressure observation data, air temperature data, and weather satellite data to discover monthly average weather conditions semi‐objectively. The results obtained, will enable this research to analyze climate change in order to provide early prevention efforts. Keywords: Biodiversity, Carbon, Climate, Weather, Hyperspectral.

1. Introduction

Indonesia is developing country located in Asian continent and situated right at the equatorial line. Having such geographical condition, Indonesian certainly has a very large area of forest. One location in Indonesia with large forest area is in Southeast Sulawesi and it can be categorized into tropical, mangrove and savanna forest. Tropical forest is inhabited by more than of total species of plants in the world Whitmore. , Foody. , Thomas et al . owever, detailed spatial Derick Christopher Ambo Masse. Tel.: + ‐8 ‐ ‐ . E ‐mail address:derickchristopher.amsgmail.com 409 information about composition and distribution of tree and its fauna diversity is still limited to some areas Clark et al. . Similarly, both mangrove and savanna forests occupy large distribution area and have their own roles as well. For example, mangrove plays important role on coast productivity Analudin. . For remote sensing scientist, satellite data can be used to assess biodiversity Lucas et al. and to contribute on dead and living carbon assessment Asner and eibretch, . Based on those facts, the main objective of this research is to analyze climate change based on biomass reserve and plant diversity in South east Sulawesi. n particular, this research is aimed: to estimate biomass carbon existence to find out what climate change will happen by analyzing biomass existence in Southeast Sulawesi forests.

2. Material and Methods

2.1. Location Location of this research is the whole region of Southeast Sulawesi Figure with astronomic position is between ° ’‐ ° ’ South Latitude and ° ’‐ ° ’ East Longitude Biro umas. . Figure . Southeast Sulawesi Map. 410 2.2. Image Processing First stage of image processing of Landsat 8 is geometric and radiometric correction. mage correction is a conditioning operation so that the image to be used can really provide accurate information geometrically and radiometrically Danoedoro. . After being corrected, the image will be analyzed using Normalized Difference Vegetation ndex NDV formula to get vegetation density point Jensen. . Next stage is a classification process to classify and separate each areas based on their own ecosystem vegetation types. 2.3. NDVI Analysis Vegetation index can be used to measure biophysics parameter such as biomass, chlorophyll, LA Leaf Area ndex , etc. Jensen. . To separate vegetation and non vegetation on satellite imaging, normalized difference vegetation index method is used Gougeon and Leckie. , Danoedoro. , with equation as follows: Next, correlation analysis will be conducted to seek for relationship between vegetation density and area use. At this step, the new analysed image will be produced and then classified into tropical, mangrove or savannah forest. n order to interpret the relationship between two variables, the r elationship power correlation criteria Table is applied Rahmi. . Table . Relationship power correlation criteria Score Correlation relationship No correlation . – . Very weak correlation . – . Medium correlation . – . Strong correlation . – . Very strong correlation Perfect correlation 2.4. Biomass and Carbon Stock Estimation Forest carbon total reserve measurement is based on biomass total content measurement and organic stuff at carbon pool PCC. . Carbon biomass estimation on land surface is calculated by building up allometric equation ministry of forestry. . Applied allometric equation is: To calculate tree volume, geometric equation is used by considering tree height and diameter data as follows: Carbon stock estimation based on biomass needs biomass conversion factor value which is called as carbon fraction forestry ministry. , formulated as: Carbon fraction value applied in Equation is . PCC. .