• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.150-150
• /
• 2022

This study investigates the roles of decomposition methods on high accuracy in daily rainfall prediction from light gradient boosting machine (LightGBM) algorithm. Here, empirical mode decomposition (EMD) and singular spectrum analysis (SSA) methods were considered to decompose and reconstruct input time series into trend terms, fluctuating terms, and noise components. The decomposed time series from EMD and SSA methods were used as input data for LightGBM algorithm in two hybrid models, including empirical mode-based light gradient boosting machine (EMDGBM) and singular spectrum analysis-based light gradient boosting machine (SSAGBM), respectively. A total of four parameters (i.e., temperature, humidity, wind speed, and rainfall) at a daily scale from 2003 to 2017 is used as input data for daily rainfall prediction. As results from statistical performance indicators, it indicates that the SSAGBM model shows a better performance than the EMDGBM model and the original LightGBM algorithm with no decomposition methods. It represents that the accuracy of LightGBM algorithm in rainfall prediction was improved with the SSA method when using multivariate dataset.

• Journal of Environmental Science International
• /
• v.28 no.1
• /
• pp.125-135
• /
• 2019

For the purposes of enhancing usability of Numerical Weather Prediction (NWP), the quantitative precipitation prediction scheme by machine learning has been proposed. In this study, heavy rainfall was corrected for by utilizing rainfall predictors from LENS and Radar from 2017 to 2018, as well as machine learning tools LightGBM and XGBoost. The results were analyzed using Mean Absolute Error (MAE), Normalized Peak Error (NPE), and Peak Timing Error (PTE) for rainfall corrected through machine learning. Machine learning results (i.e. using LightGBM and XGBoost) showed improvements in the overall correction of rainfall and maximum rainfall compared to LENS. For example, the MAE of case 5 was found to be 24.252 using LENS, 11.564 using LightGBM, and 11.693 using XGBoost, showing excellent error improvement in machine learning results. This rainfall correction technique can provide hydrologically meaningful rainfall information such as predictions of flooding. Future research on the interpretation of various hydrologic processes using machine learning is necessary.

• Atmosphere
• /
• v.26 no.1
• /
• pp.141-158
• /
• 2016

The objective of this study is to examine the impact of urban canopy and the horizontal resolution on simulated meteorological variables such as 10-m wind speed, 2-m temperature and precipitation using WRF model for a local, convective rainfall case. We performed four sensitivity tests by varying the use of urban canopy model (UCM) and the horizontal resolution, then compared the model results with observations of AWS network. The focus of our study is over the Seoul metropolitan area for a convective rainfall that occurred on 16 August 16 2015. The analysis shows that mean diurnal variation of temperature is better simulated by the model runs with UCM before the convective rainfall. However, after rainfall, model shows significant difference in air temperature among sensitivity tests depending on the simulated rainfall amount. The rainfall amount is significantly underestimated in 0.5 km resolution model run compared to 1.5 km resolution, particularly over the urban areas. This is due to earlier occurrence of light rainfall in 0.5 km resolution model. Earlier light rainfall in the afternoon eliminates convective instability significantly, which prevents occurrence of rainfall later in the evening. The use of UCM results in a higher maximum rainfall in the domain, which is due to higher temperature in model runs with urban canopy. Earlier occurrence of rainfall in 0.5 km resolution model is related to rapid growth of PBL. Enhanced mixing and higher temperature result in rapid growth of PBL, which provides more favorable conditions for convection in the 0.5 km resolution run with urban canopy. All sensitivity tests show dry bias, which also contributes to the occurrence of light precipitation throughout the simulation period.

• Journal of Korea Water Resources Association
• /
• v.50 no.1
• /
• pp.29-35
• /
• 2017

A vehicle rainfall sensor is made to control the operating speed of wipers depending on rainfall. Therefore this is the apparatus to determine the velocity phase of the wipers roughly based on the amount of rainfall. However, the technology which can judge the size of rainfall amount besides determining speed level of the wipers is developing according to the development of the function of rainfall sensor due to the development of technology. In this study, a rainfall measurement by using light scattering by precipitation particles was used. This measurement is to use light signal reflection from front glass and the bigger particle is the less detection of light by light scattering. The detection area of the rainfall sensor and detection channel were extended sizes to increase the accuracy of the rainfall. Also the W-S-R relational expression was developed by using a relationship between the specific precipitation (R) and the amount of sensor detection (S) when there is speed change of the wipers (W) and an indoor rainfall apparatus was used to convert sensing signal to rainfall. The signal system of vehicle rainfall sensor can be converted to the actual rainfall amount by using this formula and if this is provided to users then the vehicle observation network can produce higher-resolution than actual observation network can be produced.

• The Journal of Engineering Geology
• /
• v.24 no.4
• /
• pp.463-471
• /
• 2014

This paper considers the effect of rainfall on non-point source (NPS) pollutant loads. The impact of runoff on the occurrence of NPS pollutants was found to be influenced by rainfall amount, rainfall intensity, and the number of antecedent dry days (ADD), both independently and in combination. The close correlation ($R^2$ = 0.9920) between rainfall and runoff amounts was demonstrated at the study site (a flower farm) over the period between January 2011 and December 2013. The relationships among pollutant levels, runoff, and rainfall was not satisfactory results except for the Biochemical Oxygen Demand ($BOD_5$). The correlation coefficients between $BOD_5$, and both runoff and rainfall, were greater than 0.92. However, the relationships of other pollutants, such as Suspended Solid (SS), Chemical Oxygen Demand ($COD_{Mn}$), Total Nitrogen (TN), and Total Phosphorus (TP), with runoff and rainfall had correlation coefficients of less than 0.70. The roles of rainfall was different from rainfall categories on the occurrence of runoff. Instantaneous rainfall intensity was a principle factor on the occurrence of runoff following light rainfall events (total ${\leq}30mm$). For rainfall of intermediate intensity (total precipitation 31-50 mm), the combined effect of both average rainfall intensity and ADD was found to influence runoff generation. We conclude that the control of NPS pollutants with the reflection of the climate change that makes the remarkable effect of amounts and forms on the rainfall and runoff.

• Journal of the Korean Wood Science and Technology
• /
• v.47 no.3
• /
• pp.311-323
• /
• 2019

The number of newly constructed traditional Korean houses, i.e., Hanoks, and light-frame buildings is increasing. However, related research is limited owing to the lack of awareness regarding safety evaluations. Therefore, this study conducted an outdoor exposure test to accurately evaluate wooden constructions. Spruce, pine, and fir (SPF) material was monitored for a year, wherein the SPF material was artificially dried under 18% moisture content, and its physical properties and color differences were measured once a month. Large differences were observed in the material's weight and moisture content, which are indexes sensitive to daily range and rainfall; however, no significant difference was found for other basic properties in the pre and post test results. Herein, $L^*$, $a^*$, and $b^*$ values represent color differences; these values exhibited a general decrease after the test. Such differences were attributed to the loss of lignin in the wood. The color difference value was high between the months of May and July, when the daily range and rainfall significantly fluctuated. Multiple regression analysis was performed on the $a^*$ value (redness indicator), daily range, rainfall, and ultraviolet index. The results indicated that the daily range influenced redness the most. According to the estimated regression equation, the daily range and redness are positively correlated. Based on the results, the types and influence of independent variables on color difference are expected to change as the wood's duration of outdoor exposure and the amount of data obtained both increase.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.153-153
• /
• 2022

Large-scale and accurate observations at fine spatial resolution through a means of remote sensing offer an effective tool for capturing rainfall variability over the traditional rain gauges and weather radars. Although satellite rainfall products (SRPs) derived using two major estimation approaches were evaluated worldwide, their practical applications suffered from limitations. In particular, the traditional top-down SRPs (e.g., IMERG), which are based on direct estimation of rain rate from microwave satellite observations, are mainly restricted with their coarse spatial resolution, while applications of the bottom-up approach, which allows backward estimation of rainfall from soil moisture signals, to novel high spatial resolution soil moisture satellite sensors over South Korea are not introduced. Thus, this study aims to evaluate the performances of a state-of-the-art bottom-up SRP (the self-calibrated SM2RAIN model) applied to the C-band SAR Sentinel-1, a statistically downscaled version of the conventional top-down IMERG SRP, and their integration for a targeted high spatial resolution of 0.01° (~ 1-km) over central South Korea, where the differences in climate zones (coastal region vs. mainland region) and vegetation covers (croplands vs. mixed forests) are highlighted. The results indicated that each single SRP can provide plus points in distinct climatic and vegetated conditions, while their drawbacks have existed. Superior performance was obtained by merging these individual SRPs, providing preliminary results on a complementary high spatial resolution SRP over central South Korea. This study results shed light on the further development of integration framework and a complementary high spatial resolution rainfall product from multi-satellite sensors as well as multi-observing systems (integrated gauge-radar-satellite) extending for entire South Korea, toward the demands for urban hydrology and microscale agriculture.

• Asian Journal of Atmospheric Environment
• /
• v.8 no.4
• /
• pp.202-211
• /
• 2014

In this study, long-term rainfall data with irregular spatial distribution in Seoul, Korea, were separated into individual precipitation events by the inter-event time definition of 6 hours. Precipitation washout of $PM_{10}$ and $NO_2$ concentrations in the air considering various complex factors were analyzed quantitatively. Concentrations of $PM_{10}$ and $NO_2$ in the atmosphere were lower under condition of rainfall compared to that of non-precipitation, and a noticeable difference in average $PM_{10}$ concentrations was observed. The reduction of concentrations of $PM_{10}$ and $NO_2$ by rainfall monitored at road-side air monitoring sites was also lower than that of urban air monitoring sites due to continuous pollutant emissions by transportation sources. Meanwhile, a relatively smaller reduction of average $PM_{10}$ concentration in the atmosphere was observed under conditions of light rainfall below 1 mm, presumably because the impact of pollutant emission was higher than that of precipitation scavenging effect, whereas an obvious reduction of pollutants was shown under conditions of rainfall greater than 1 mm. A log-shaped regression equation was most suitable for the expression of pollutant reduction by precipitation amount. In urban areas, a lower correlation between precipitation and reduction of $NO_2$ concentration was also observed due to the mobile emission effect.

• Korean Journal of Ecology and Environment
• /
• v.42 no.4
• /
• pp.487-494
• /
• 2009

We compared spatial and temporal variations of water chemistry between high-flow year ($HF_y$) and low-flow year ($LF_y$) in an artificial lentic ecosystem of Daechung Reservoir. The differences in the rainfall distributions explained the variation of the annual inflow and determined flow characteristics and water residence time and modified chemical and biological conditions, based on TP, suspended solids, and chlorophylla, resulting in changes of ecological functions. The intense rainfall and inflow from the watershed resulted in partial disruption of thermal structure in the metalimnion depth, ionic dilution, high TP, and high suspended solids. This condition produced a reduced chlorophyll-a in the headwaters due to low light availability and rapid flushing. In contrast, reduced inflow and low rainfall by drought resulted in strong thermal difference between the epilimnion and hypolimnion, low inorganic solids, high total dissolved solids, and low phosphorus in the ambient water. The riverine conditions dominated the hydrology in the monsoon of $HF_y$ and lacustrine conditions dominated in the $HF_y$. Overall data suggest that effective managements of the flow from the watershed may have an important role in the eutrophication processes.

• Journal of Sensor Science and Technology
• /
• v.23 no.4
• /
• pp.259-265
• /
• 2014

When raining or snowing, windshield wiper system is very important for safety of driver. However, manual wiper system frequently needed to be controlled for sufficient visibility and it was very uncomfortable. So, rain sensor which controls automatically was developed. This rain sensor technology uses optical sensing technique sensed the rainfall by receiving reflected light of rain dropped on the windshield. The technology used optical sensor was simple and easy to implement as a rain sensing system in the car. However, it is sometime shown low accuracy to measure rainfall on the windshield when affected by ambient lights from surroundings. It is also given inconvenience to the driver to control the car. To solving these problems, we propose a rain sensing system using image sensor and the fuzzy wiper control algorithm.