GROUND-BASED CLOUD OBSERVATION FOR SATELLITE-BASED CLOUD DISCRIMINATION AND ITS VALIDATION

Solar radiation is a source of energy for ground surfaces. Clouds control energy budget on both local and global scales. We propose the simple ground-based observation method for various sky conditions instead of the existing visual observation. In the paper, we mainly describe how to discriminate clouds area on whole sky images, how to overlay the cloud areas to the satellite-based images and the test experiment. As for the discrimination of sky conditions, we use Sky Index (SI) and Brightness Index (BI) calculated from whole sky imageries. SI shows the extent of the blueness and grayscale and BI indicates the extent of the brightness. Sun, cloud and blue sky areas are divided by SI and BI threshold. As an extension of ground-based observation to satellite-based observation, clouds area discriminated by ground-based observation overlay to satellite images. Through this experiment, the utilities of ground-based cloud observation are discussed. The proposed ground-based cloud observation method and its satellite-based extension should be connected to satellite products with high quality by implementation of sky and clouds observation covering a wide area.


INTRODUCTION
Clouds control energy budget on both local and global scales.Generally, it is well known that lower clouds cool the surface and upper clouds have the effects to warm the surface (Clement et al, 2009).Cumulus clouds especially in atmospheric boundary layer play crucial role in modulating the exchange of radiation, heat and moisture (Freedman et al. 2001).The size and the lifespan of cumulus clouds are small and short, therefore it is difficult to observe and monitor them spatially and temporally.Satellite-based observation and its products related to clouds properties have been useful to estimate some physical values such incident radiation/ Photosynthetic Active Radiation (PAR) and Fraction of PAR etc., and to be advanced various high quality satellite products.However, there are still some uncertainness such as cloud motion, cloud coverage, cloud type and these effects by satellite-based observation.Therefore, the efficient ground-based cloud observation is necessary to validate satellite-based observation such existing Terra & Aqua/ Moderate Resolution Imaging Spectroradiometer (MODIS) and under planning Global Change Observation Mission -Climate (GCOM-C) / Second generation GLobal Imager (SGLI) etc.Recently, the ground-based cloud observation systems have been developed by using whole sky camera in world widely instead of the existing visual observation (Calbo and Sabburg, 2008;Yamashita and Yoshimura, 2008;Heinle et al., 2010).The purpose of this study is to discriminate the cloud and overlay the cloud areas with satellite-based images by the efficient and simple ground-based cloud observation using commercial based digital camera and fish eye lens in order to validate the satellite-based observation.In the paper, we mainly describe how to discriminate the cloud area on whole sky images, how to overlay the cloud areas to the satellite-based images and the test experiment.Through this experiment, the utilities of ground-based cloud observation are discussed.

Camera system
Automatic-capturing Digital Fisheye Camera (ADFC) is the instruments to take and accumulate the whole sky imageries.ADFC consists of digital camera (Nikon, Coolpix4500) with fisheye lens (Nikon, Fisheye Converter FC-E8), waterproof hard case and remote controlled cable connected to PC (Tsuchida et al, 2005).In order to decrease the influence for CCD saturation caused by the strong sun light, the neutral density filter with 10% transparency (FUJIFILM, ND1.0) puts between the digital camera and the fisheye lens.The camera is set up to take imageries at every two-minute interval by fixed exposure of the aperture as F2.6 and shutter speed as 1/500s.The image size has 2204pixels by 1704lines with RGB colors and JPEG (1/4 compressed) format instead of RAW or Tiff format to accumulate the large volume of image files.In this study, we already have confirmed that the almost no influences between JPEG (1/4 compressed) and RAW/Tiff formats after image processing.Also, digital number (DN) of RGB channels can be used instead of radiance by fixing aperture and exposure.The head of digital camera of ADFC is fixed to north direction and levelled to horizon.

Discrimination of cloud, blue sky and sun
The sky conditions consist of each state of sun, clouds and blue sky.In order to discriminate the sky conditions from whole sky imageries, we use two indices which show the levels of the blueness/greyscale and the brightness calculated from the imageries.Here, the method how to discriminate the sky condition is described.Figure 1 shows the flowchart to classify cloud and blue sky areas from the whole sky imagery taken at the time (t).
In order to discriminate cloud and blue sky areas from the whole sky imageries, we use Sky Index (SI) and Brightness Index (BI) which are calculated from RGB channels of the image.SI shows the blueness/greyscale and BI shows the brightness in pixels.SI and BI are expressed by equation ( 1) and ( 2). (1) where DN B ; digital number of Blue channel DN G ; digital number of Green channel DN R ; digital number of Red channel n ; bit number of quantization level SI values have the range between -1.0 and 1.0.The blue sky area in the RGB image has the higher digital number in blue channel and the lower digital number in red channel.On the other hand, the clouds show white or grey on the image.Thus, the higher value of SI shows bluer sky, the value of near zero shows the clouds and the sun.BI values are expressed by the range between 0 and 1.0.The bright pixels on the image show high BI.The pixels in the sun area have BI = 1 and SI = 0.The whole sky areas consisted of the sun, clouds and blue sky are classified by the threshold curve which was derived from two dimensions of SI and BI coordinates.The threshold curve for sky area classification was determined by counting the frequency of SI and BI combination's pixel.In this case, the threshold curve was used as BI = e -9*SI .When the coordinates plot under the threshold curve, these are classified as cloud area.
The coordinates upper the curve are classified as blue sky area.The sun area in the imagery has the coordinates of SI = 0 and BI = 1.In this way, the sky classified image was generated by SI & BI threshold.

OVERLAYING TO SATELLITE IMAGE
As for the method of overlaying with cloud areas of both ground and satellite-based images, the cloud area on whole sky image are projected to ground surface plane with map coordinate.Our used fisheye lens has equidistant projection.In case of setting the ADFC fixed to north direction and levelled horizontally, the central pixel of whole sky (hemisphere) image corresponds to the zenith.According to equidistant projection, the distance from the image centre is in proportion to zenith angle shown in equation ( 3).Consequently, the zenith () and azimuth () angles on whole sky image are able to calculate from the image coordinates (u, v).And then, the horizontal distance (HD) from the groundbased observation point, that is central position of whole sky image, to the target cloud can be estimated by given the approximate cloud height supposed from each cloud type, and the cloud areas on whole sky image are projected to map coordinates (x, y) using ,  and HD.
In these calculation and conversion, any distortions of fisheye lens and the curvature of ground surface are left out of considerations.

Calculation of zenith and azimuth angles
Calculation of zenith and azimuth angles at any hemisphere image coordinates can be shown in equation ( 4) and (5).
In this study, the image size of hemispherical area has 1650 pixels and 1650 lines.Therefore the radius of circle (R) is 825 pixels and the image centre coordinate (u o , v o ) is (824, 824) in case of the image upper left coordinates as (0, 0).

Cloud area conversion to map coordinates
For each pixel classified as cloud area on whole sky image, the projected cloud area on the ground can be converted to map  coordinates using zenith () and azimuth () angles by given any cloud heights approximately.Generally, lower clouds such stratocumulus appears from near the ground to 2km and in case of cumulus, it appears from 0.6 to 6km.The conversion from the each pixel (u i , v i ) of cloud area on whole sky image to map coordinates (x i , y i ) can be shown in equation ( 6) and ( 7).

RESULT AND DISCUSSSION
As one of experiment, we examine to compare and overlay with several whole-sky images of different cloud types and those synchronous Terra & Aqua/ MODIS images observed in central Kyoto, Japan.The grand-based cloud observation point is located at (N35.018243°, E135.768278°).In this experiment, we use UTM (Zone53N, WGS84) as map projection.
Figure 2 shows one of the results of cloud discrimination from whole sky images taken at 12:54 and 12:56 on Aug. 26 th 2005.
Red circle corresponds to the area of zenith angle less than 65 degrees.Cloud type is recognized as cumulus of small size visually.Also cloud type such cumulus and stratocumulus have low SI, very close to SI = 0, and BI values digitally.
Outside of this red circle seems to overestimate cloud coverage because of the long pass of atmospheric layer from the view at large zenith angle.Here, the target area is in less than 65 degrees of zenith angle.
26 th 2005 as one of the sample satellite images in this experiment.
There are not seen so much cloud coverage around the observation point by comparing with northern part of this MODIS image.However, it is certainly confirmed the clouds existence in the red circle visually.Also, there is the possibility of this study on the assumption that this ground-based clouds observation method is applied across a wide area.For the validation of satellite-based clouds and earth surface observation, it seems that this observation system is suitable to set up at many sites, and also useful to collaborate with the existing ground-based observation networks such Asia Flux, AERONET and SKYNET etc.The proposed observation method and its applications should be connected to satellite products with high quality by implementation of sky and clouds observation covering a wide area.

CONCLUSION
We proposed the simple ground-based observation method to discriminate clouds on whole sky image and one of extensions to the satellite-based observation by overlaying both ground and satellite-based images.
Through this study, we confirmed that our developed groundbased cloud observation method is useful to discriminate cloud existence and area of cloud coverage, and to overlay with satellite-based cloud image.Also, this ground-based cloud observation can be expanded into wide area and set many observation points for application to various satellite images including geostationary satellites with very high temporal resolution.
distance from zenith (image centre) R; Radius of whole sky image area (pixels) ; Zenith angle (degrees) v i ; Image coordinates of column and row (pixels) u o , v o ; Image centre coordinates (pixels)

Figure 1 .
Figure 1.The flowchart of cloud, blue sky and sun discrimination by whole sky imageries Figure 2. Results of sky classified image (right) from whole sky images (left) of 2-minutes different images.Cloud discrimination area is painted as light green on whole sky image.Red circle shows the area of zenith angle less than 65 degrees.

Figure 3 .
Figure 3. Aqua/ MODIS VIS (ch.1) and NIR (ch.2) composite image (RGB: ch.1,2,1) with 250m ground resolution observed at 12:55(JST) on Aug. 26 th , 2005.Red x is the point of ground-based cloud observation.Red circle indicates the overlaying area with zenith angle less than 65 degrees in case of HD calculation as CH apr = 2000m.