Chang-Suk Lee

Evapotranspiration in Korea estimated by application of a neural network to satellite images

Previous biophysical and empirical models of evapotranspiration retrieval are difficult to parameterize because of the effects of the nonlinear biophysics of plants, terrestrial and solar radiation and soils, despite attempts made using various satellite products. In this study, the multilayer feed-forward neural network approach with Levenberg–Marquardt back propagation (LM-BP) was used to successfully estimate evapotranspiration using the input of various satellite-based products. When applying neural network training, value-added satellite-based products such as normalized difference vegetation index (NDVI), normalized difference water index (NDWI), land surface temperature (LST), air temperature and insolation are used instead of only spectral information from satellite sensors to reflect the spatial representativeness of the neural network. The evapotranspiration estimated from the neural network with input parameters showed better statistical accuracy than the MODIS products (MOD16) and Priestley–Taylor methods when compared with ground station eddy flux measurements, which were considered as reference data. Additionally, the temporal variation in neural network evapotranspiration well reflected seasonal patterns of eddy flux evapotranspiration, especially for the high cloudiness in the summer season.

Estimating midday near-surface air temperature by weighted consideration of surface and atmospheric moisture conditions using COMS and SPOT satellite data

The measurement of near-surface air temperature (Ta) is critically important for understanding the Earth’s energy and water circulation system and for diverse modelling applications. Ta data obtained from meteological ground stations are basically available but not suitable for large-scale areas, because of their spatial limitation. Remote-sensing techniques can provide a spatially well-distributed Ta map. However, the current remote-sensing methodology for Ta mapping has accuracy inferior to common expectations in terms of the region of various terrestrial ecosystems and climatic conditions. Our aim was to develop a midday Ta retrieval algorithm with reasonable accuracy over Northeast Asia during one seasonal year. In consideration of the various environmental conditions in our study area, Ta was calculated using land surface temperature and the normalized difference vegetation index in the nine cases derived from the combination of surface and atmospheric moisture conditions, and a weighting factor was applied to reduce the bias error among Ta results from nine cases. The reasonable pixel window size was established as 13 × 13. The validation process yielded a coefficient of determination (R2), root mean square error, and bias values of 0.9401, 2.8865 K, and 0.4920 K, respectively. Although the study area includes diverse land-cover and climatic conditions, our satellite-derived Ta data provided better results compared with a previous study of only four cases with no weighting function in the Korean peninsula. Our suggested methodology will be useful in estimating Ta using satellite data, particularly over complex land surfaces.

Updating Absolute Radiometric Characteristics for KOMPSAT-3 and KOMPSAT-3A Multispectral Imaging Sensors Using Well-Characterized Pseudo-Invariant Tarps and Microtops II

Radiometric calibration of satellite imaging sensors should be performed periodically to account for the effect of sensor degradation in the space environment on image accuracy. In this study, we performed vicarious radiometric calibrations (relying on in situ data) of multispectral imaging sensors on the Korea multi-purpose satellite-3 and -3A (KOMPSAT-3 and -3A) to adjust the existing radiometric conversion coefficients according to time delay integration (TDI) adjustments and sensor degradation over time. The Second Simulation of a Satellite Signal in the Solar Spectrum (6S) radiative transfer model was used to obtain theoretical top of atmosphere radiances for both satellites. As input parameters for the 6S model, surface reflectance values of well-characterized pseudo-invariant tarps were measured using dual ASD FieldSpec® 3 hyperspectral radiometers, and atmospheric conditions were measured using Microtops II® Sunphotometer and Ozonometer. We updated the digital number (DN) of the radiance coefficients of the satellites; these had been used to calibrate the sensors during in-orbit test periods in 2013 and 2015. The coefficients of determination, R2, values between observed DNs of the sensors, and simulated radiances for the tarps were more than 0.999. The calibration errors were approximately 5.7% based on manifested error sources. We expect that the updated coefficients will be an important reference for KOMPSAT-3 and -3A users.

Improved estimation of insolation by using calibrated COMS MI images over South Korea

In this study, a correction function based on a comprehensive vicarious calibration was applied to Communication, Ocean and Meteorological Satellite (COMS) and Meteorological Imager (MI) visible channel for improving the estimate of insolation. The COMS MI visible channel was calibrated by using natural targets such as ocean, desert, water cloud and deep convective cloud, which are carefully selected due to their theoretical reference reflectance. Based on calibration with four targets, the corrected COMS MI visible channel is used as an input parameter of an insolation physical model, especially for interpreting cloud effects. Results showed that corrected insolation has better accuracy than previous methods for cloudy conditions. Seasonal bias of insolation, which generally has the worst accuracy in a cloudy summer period, is improved by using calibrated COMS MI visible data.

Effects of GSICS correction on estimation of sea surface temperature using COMS data

A main objective of the Global Space-Based Inter-Calibration System (GSICS) is to ensure consistent accuracy of thermal infrared (IR) channels of satellite measurements. In this article, IR channels observed by the Communication, Ocean, and Meteorological Satellite (COMS) were corrected using observations by the Infrared Atmospheric Sounding Interferometer (IASI) on board Meteorological Operation (MetOp) satellites as a reference. We then produced sea surface temperature (SST) products using GSICS-corrected COMS/Meteorological Imager (MI) IR data and evaluated the accuracies of the SST data by comparing with buoy data. Our results showed that SST produced from GSICS-corrected IR data was more accurate than that from original IR data, with a difference of 0.01–0.2ºC. Differences in SST between the original and GSICS-corrected IR showed a linear proportional relationship around IR2 (12.0 μm) during the day and IR1 (10.8 μm) at night. In the IR1 analysis, accuracies of daytime SST were improved significantly when the GSICS correction was applied to COMS IR data; also, the mean and median values of SST discrepancies (between the original and GSICS-corrected IR) according to changes in brightness temperature difference (BTD) were close to 0ºC. We found that the accuracies of SST estimation from satellite IR bands were greatly improved by applying the GSICS correction.

11-year variability of summer snow cover extent over Himalayas

Snow is a component of the cryosphere which has played an important role in Earth energy balance. Northern hemisphere snow cover extent (SCE) has steadily decreased since 1980 and in recently the trend of SCE is sharply decreased. Because Himalaya regions shows most significant changes except for the Arctic, we analyzed this region for SCE. We used Moderate Resolution Imaging Spectroradiometer (MODIS) snow product from 2001 to 2011 in august. Analysis was made by considering some conditions (region, elevation, longitude and climate) which can affect the changes in SCE. The entire SCE in Himalaya for 11 years has steadily increased(+55,098 km2). Trends for SCE in western region has increased(+77,781km2), But trend for central and eastern have decreased -3,453 km2, -19,230km2, respectively. According to elevation increases, the ratio of snow in each study area is increased. In 30°N~35°N SCE shows increased trend, 27°N~28°N shows decreased trend. In tundra climate, trends for SCE are similar to regional analysis. whereas the result in tropical climates trend was increased. these performed result shows different side for change of SCE depending on each condition. The result of this study were similar to the rapid decline of the northern hemisphere SCE area in recent. The result of this study can be used to help management to water budget in Central-Asia country located to Himalayas.

Estimation of maximum air temperature using COMS data in Northeast Asia

Air temperature (Ta) plays important role for the circulation of energy and water between the surface and atmosphere. Ta was accurately measured from ground observation stations. However, the number of ground observation stations is limited, and Ta is influenced from temporal and spatial change. In this study, Ta was estimated using satellite data from April 2011 to March 2012 in the Northeast Asia where consist of the various ecosystem. States of surface and atmosphere were considered through Normalized Difference Water Index (NDWI) and the differences of brightness temperature values of 11μm (TBB1) and 12μm (TBB2). Dataset was divided into nine cases that had seasonal characteristics according surface states (NDWI) and atmosphere states (TBB1-TBB2). Ta was acquired from 174 ground observation stations, and multiple regression equation of each case was consisted of LST, NDVI, TBB1-TBB2. The weighting region was set to be within 8.33% of total density from boundary area of cases in order to reduce the errors that can occur due to the small value. The weighting was applied as distance from the nearest four points. The spatial representativeness of estimated Ta was determined as 9 by 9 window size. R-squared of estimated Ta from satellite was 0.94, RMSE was 2.98 K, Bias was 0.56 K.

An influence assessment of GSICS correction using sea surface temperature from geostationary satellite: COMS

There is a strong need for accurate estimation of radiance from satellite regarding establishing a climate records such as global climate circulation, change and Earth’s atmosphere. It is important that exact radiance measurements from satellite to numerical weather prediction models for climate change detection. Furthermore, accurate measurements from satellite rely on calibration of channel data in terms of the radiometric characteristics. Related to improved calibration and inter-calibration of the sensors, the World Meteorological Organization (WMO) and the Coordination Group for Meteorological Satellite (CGMS) initiated the Global Space-based Inter-Calibration System (GSICS) in 2005, which provide coefficients to the user community to adjust satellite observations. To assess influence of the GSICS corrections and impacts of input parameters changes on satellite products, the coefficients of the GSICS corrections were applied to infrared (IR) data from Communication Ocean and Meteorological Satellite (COMS), which have Meteorological Imager (MI) sensor for meteorological missions. The IR data centered at wavelengths of 10.8 (IR1) and 12.0μm (IR2) from the COMS MI were compared with that of the Infrared Atmospheric Sounding Interferometer (IASI) sensor, which is reference sensor of the GSICS corrections. The IR1 and IR2 data that were corrected by GSICS produced Sea Surface Temperature (SST), which has been influenced by input parameters such as IR data and solar zenith angle. As a result of comparison with in situ measurements, the Global Telecommunication System (GTS) buoy data, COMS IR data that were corrected by the GSICS corrections produced high quality products of SST than original COMS IR data.