|
|
|
Oceanography / Meteorology
|
Mid-day Atmospheric Humidity from Thermal Infrared Observation of the NOAA-14 AVHRR Satellite: Validation in Tropical Environment
T. Sarvanaapavan*, D.G. Dye**, and R. Shibasaki*
*Global Engineering Laboratory
Department of Civil Engineering
Institute of Industrial Science
University of Tokyo, Japan
**Center of Remote Sensing
Department of Geography
Boston University
Massachusetts 02215, USA
Abstract
Recent research introduced a method of estimating mid-day atmospheric humidity based on the "split-window" technique that takes advantages of the attenuation in thermal anfra-red (TIR) measurements by atmospheric water vapor. The method has been yet been evaluated in low-latitude tropical environments which have energy, temperature and precipitation regimes that differ substantially from than those of temperature latitudes.
We examine the effectiveness of the method to estimate mid-day atmospheric humidity through an application at selected sample sites in the seasonally moist tropical environment of Thailand. We employ daily 1.1 km resolution observations of the NOAA-14 AVHRR satellite in this study. Our analysis shows substantial promise for applying in this tropical environment. We also demonstrated application of the method for generating spatial maps on atmospheric humidity.
Introduction
Satellite remote sensing involves interpretation of emitted radiation, received and measured by an earth observing sensor. The emission from the surface is determined by the atmosphere, presence of water vapor, CO2 and O3 can be neglected when compared with water vapor as they have lesser optical depth in the TIR spectral region than water vapor (Price 1983). This atmospheric attenuation's considered as a to indicate the water presence in the atmosphere (Goward et al, 1994).
The dual TIR spectral sensors on the advanced Very High Resolution Radiometer (AVHRR) (channel 4, 10.3-11.3 mm) initially were designed to evaluate atmospheric water vapor attenuation in the 10-12 mm spectral region, with the objective of producing accurate ocean surface ocean surface temperature (Prabhakar et al., 1974, McClain et al., 1985), Goward. et al. (1994) introduced a method, based on the TIR attenuation by atmospheric water, to estimate mid-day atmospheric humidity and found reasonable success in validation in the temperate environment of Oregon, USA. The method, however, has not yet been evaluated in low-latitude tropical environments which have energy, temperature latitudes. Validation of the existing techniques in contrasting climates and contrasting landscape patterns is important prior to regional, continental and global application. In this study we focus on examine the method of estimating atmospheric humidity by comparing satellite estimation with ground observations at selected sample sites in the tropical environment of Thailand.
Background
In the estimation of the atmospheric humidity, the attenuation of the TIR measurements by water vapor is used as an indicator of the amount of the water vapor in the atmospheric. Wome researchers (Dalu 1986, Jedlovec 1990) identified the correlation between TIR split window measurements and precipitalble water in the entire column of the atmosphere between the sensor and surface (approximately 800 km depth with an area of 1 km x 1 km cross section, at nadir). The split window measurement and its relation with the water vapor in the lower portion of the mid-day atmospheric, surrounding the surface meteorological stations and the vegetation canopy have been discussed by Goward et al. (1994).
It is generally not feasible to estimate atmospheric humidity because atmospheric conditions can vary between inversion conditions, thermally stable conditions, and lase conditions, thermally unstable conditions (Resenberg et al., 1983). However, water vapor is furnished to the air only from evaporation surface of land and water. Water vapor transport into and through the layer of air adjacent to the ground is analogous to heat transport. Heat and water vapor are transferred to the bulk air primarily by convection or turbulent transport. During the mid-day atmospheric conditions, when surface heating produces strong vertical mixing in the active boundary layer, the perceptible water in the near surface atmosphere becomes well mixed within the active boundary layer. The satellite over pass time, approximately 1:30 p.m., is typically coincident with lapse conditions, when the atmosphere potential for estimating atmospheric humidity at the time of satellite over passing. Goward et al. (1994) introduced an approach by which the correlation between the calculated least-squares slope relation between the AVHRR TIR radiometric temperature (T4 and T5 from an observation of 9x9 array, centered on each of the study sites) and absolute humidity could be found. This approach included three main assumptions. First, the atmospheric was always experiencing strong vertical mixing. Second, atmospheric humidity during the observations was relatively uniform, horizontally, over the 81 km2 region of the observation. Third, as in the surface temperature calculations, emissivity did not vary with locations within or between sites.
Data and methodology
Five multispectral image data sets of the NOAA-14 AVHRR daily observations were acquired from Japan's National Institute of Environmental studies (NIES), and the United States' national Climate Data Center (NCDC) . The data sets were selected to represent connective days in different seasons (e.g. hot and dry season, rainy season and cool and dry season of Thailand (Moncharoen et al, 1987). The AVHRR data acquired for this study were in level b format, which is raw data that have been quality controlled, assembled into discrete data sets, and to which Earth location and calibration information have been appended, but not applied (Kidwell, 1995). The pre-processing we carried out includes mapping, radiometric calibration and cloud screening.
Ground measured methodological observations were acquired for validating the satellite estimation. The source of ground observed data used in this study was the Meteorological Department of Thailand. The data set includes daily humidity and temperature measurement in three hours interval of 71 stations in Thailand. The climate sample sites (Table.1) were selected by considering the representation of different landscape and climate patterns of Thailand and the availability of cloud-free satellite observations for the sites.
Table. 1. Selected sample sites and the information on geographic location.
| Site No. | Site Name. | Latitude. | Longitude. | Elevation (m) |
| 1 | Mae Hong Son | 19° 18' N | 097° 50' E | 269 |
| 2 | Chiang Rai | 19° 55' N | 099° 50' E | 395 |
| 3 | Chiang Mai | 18° 47' N | 098° 59' E | 314 |
| 4 | Nong Khai | 17° 52' N | 102° 43' E | 175 |
| 5 | Mae Sot | 16° 40' N | 098° 33' E | 197 |
| 6 | Mukdahan | 16° 32' N | 104° 43' E | 139 |
| 7 | Chaiyaphum | 15° 48' N | 102° 02' E | 183 |
| 8 | Ubon Ratchathani | 15° 15' N | 104° 52' E | 127 |
| 9 | Supan Buri | 14° 28' N | 100° 08' E | 8 |
| 10 | Korat | 14° 58' N | 102° 05' E | 188 |
| 11 | Surin | 14° 53' N | 103° 30' E | 147 |
| 12 | Bangkok | 13° 44' N | 100° 34' E | 20 |
| 13 | Rayong | 13° 38' N | 101° 21' E | 5 |
| 14 | Chanthaburi | 12° 36' N | 102° 07' E | 4 |
|
|
|
|
|
|
|
