• Journal of Korea Water Resources Association
• /
• v.41 no.7
• /
• pp.747-759
• /
• 2008

This study, a point rainfall process model, which could represent appropriately observed rainfall data, was to select. The point process models-rectangular pulses Poisson process model(RPPM), Neyman-Scott rectangular pulses Poisson process model(NS-RPPM), and modified Neyman-Scott rectangular pulses Poisson process model(modified NS-RPPM)-all based on Poisson process were considered as possible rainfall models, whose statistical analyses were performed with their simulation rainfall data. As results, simulated rainfall data using the NS-RPPM and the modified NS-RPPM represent appropriately statistics of observed data for several aggregation levels. Also, simulated rainfall data using the modified NS-RPPM shows similar characteristics of rainfall occurrence to the observed rainfall data. Especially, the modified NS-RPPM reproduces high-intensity rainfall events that contribute largely to occurrence of natural harzard such as flood and landslides most similarly. Also, the modified NS-RPPM shows the best results with respect to the total rainfall amount, duration, and inter-event time. In conclusions, the modified NS-RPPM was found to be the most appropriate model for the long-term simulation of rainfall.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2004.05b
• /
• pp.1290-1294
• /
• 2004

In this study, new stochastic point rainfall models which can consider the correlation structure between rainfall intensity and duration are developed. In order to consider the negative and positive correlation simultaneously, the Gumbels type-II bivariate distribution is applied, and for the cluster structure of rainfall events, the Neyman-Scott cluster point process is selected. In the theoretical point of view, it is shown that the models considering the dependent structure between rainfall intensity and duration have slightly heavier tail autocorrelation functions than the corresponding independent mode]s. Results from generating long time rainfall events show that the dependent models better reproduce historical rainfall time series than the corresponding independent models in the sense of autocorrelation structures, zero rainfall probabilities and extreme rainfall events.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.14 no.2
• /
• pp.307-316
• /
• 1994

The rainfall forecasting model of the short term is improved at the point where meterological data is not gaged. In this study, the adopted model is based on the assumptions for simulation model of rainfall process, meteorological homogeneousness, prediction and estimation of meteorological data. A Kalman Filter technique is used for rainfall forecasting. In the existing models, the equation of the model is non-linear type with regard to rainfall rate, because hydrometer size distribution (HSD) depends on rainfall intensity. The equation is linearized about rainfall rate as HSD is formulated by the function of the water storage in the cloud. And meteorological input variables are predicted by emprical model. It is applied to the storm events over Taech'ong Dam area. The results show that root mean square error between the forecasted and the observed rainfall intensity is varing from 0.3 to 1.01 mm/hr. It is suggested that the assumptions of this study be reasonable and our model is useful for the short term rainfall forecasting at the ungaged point of the meteorological data.

• Journal of Korea Water Resources Association
• /
• v.34 no.6
• /
• pp.577-586
• /
• 2001

This study proposes a theoretical methodology for deriving a rainfall intensity-duration- frequency (I-D-F) curve using a simple rectangular pulses Poisson process model. As the I-D-F curve derived by considering the model structure is dependent on the rainfall model parameters estimated using the observed first and second order statistics, it becomes less sensitive to the unusual rainfall events than that derided using the annual maxima rainfall series. This study has been applied to the rainfall data at Seoul and Inchon stations to check its applicability by comparing the two I-D-F carves from the model and the data. The results obtained are as followed. (1) As the duration becomes longer, the overlap probability increases significantly. However, its contribution to the rainfall intensity decreases a little. (2) When considering the overlap of each rainfall event, especially for large duration and return period, we could see obvious increases of rainfall intensity. This result is normal as the rainfall intensity is calculated by considering both the overlap probability and return period. Also, the overlap effect for Seoul station is fecund much higher than that for Inchon station, which is mainly due to the different overlap probabilities calculated using different rainfall model parameter sets. (3) As the rectangular pulses Poisson processes model used in this study cannot consider the clustering characteristics of rainfall, the derived I-D-F curves show less rainfall intensities than those from the annual maxima series. However, overall pattern of both I-D-F curves are found very similar, and the difference is believed to be overcome by use of a rainfall model with the clustering consideration.

• Journal of Korea Water Resources Association
• /
• v.51 no.1
• /
• pp.19-33
• /
• 2018

This study proposed a rainfall-runoff model for the purpose of real-time flood warning in urban basins. The proposed model was based on the shot noise process, which is expressed as a sum of shot noises determined independently with the peak value, decay parameter and time delay of each sub-basin. The proposed model was different from other rainfall-runoff models from the point that the runoff from each sub-basin reaches the basin outlet independently. The model parameters can be easily determined by the empirical formulas for the concentration time and storage coefficient of a basin and those of the pipe flow. The proposed model was applied to the total of three rainfall events observed at the Jungdong, Guro 1 and Daerim 2 pumping stations to evaluate its applicability. Summarizing the results is as follows. (1) The unit response function of the proposed model, different from other rainfall-runoff models, has the same shape regardless of the rainfall duration. (2) The proposed model shows a convergent shape as the calculation time interval becomes smaller. As the proposed model was proposed to be applied to urban basins, one-minute of calculation time interval would be most appropriate. (3) Application of the one-minute unit response function to the observed rainfall events showed that the simulated runoff hydrographs were very similar to those observed. This result indicates that the proposed model has a good application potential for the rainfall-runoff analysis in urban basins.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.14 no.5
• /
• pp.1129-1136
• /
• 1994

A point heavy rainfall process is physically modeled. It uses meteorological variables at the ground level as its inputs. The components of the model are parameterized based on well established observations and the previous studies of cloud physics. Particular emphasis is placed on the efficiency of accretion. So we adopt the modified skew-symmetric model for hydrometeor size distribution function that is suitable for the heavy rain cloud. The dominant parameters included in the model are estimated by the optimization technique. The rainfall intensity is predicted by the model with the medium values of estimated parameters.

• Journal of Korea Water Resources Association
• /
• v.47 no.8
• /
• pp.693-701
• /
• 2014

In this study a stochastic rainfall generation model is used to analyze the structural variability of rainfall events since it has limitations in the traditional approach of measuring rainfall variability according to different durations. The NSRPM(Neyman-Scott Rectangular Pulse Model) is a stochastic rainfall generation model using a point process with 5 model parameters which is widely used in hydrologic fields. The five model parameters have physical meaning associated with rainfall events. The model parameters were estimated using hourly rainfall data from 1973 to 2011 at Seoul stations. The variability of model parameter estimates was analyzed and compared with results of traditional analysis.

• Journal of Korea Water Resources Association
• /
• v.54 no.5
• /
• pp.301-309
• /
• 2021

In this study, we tried to improve the performance of the existing U-net-based deep learning rainfall prediction model, which can weaken the meaning of time series order. For this, ConvLSTM2D U-Net structure model considering temporal consistency of data was applied, and we evaluated accuracy of the ConvLSTM2D U-Net model using a RainNet model and an extrapolation-based advection model. In addition, we tried to improve the uncertainty in the model training process by performing learning not only with a single model but also with 10 ensemble models. The trained neural network rainfall prediction model was optimized to generate 10-minute advance prediction data using four consecutive data of the past 30 minutes from the present. The results of deep learning rainfall prediction models are difficult to identify schematically distinct differences, but with ConvLSTM2D U-Net, the magnitude of the prediction error is the smallest and the location of rainfall is relatively accurate. In particular, the ensemble ConvLSTM2D U-Net showed high CSI, low MAE, and a narrow error range, and predicted rainfall more accurately and stable prediction performance than other models. However, the prediction performance for a specific point was very low compared to the prediction performance for the entire area, and the deep learning rainfall prediction model also had limitations. Through this study, it was confirmed that the ConvLSTM2D U-Net neural network structure to account for the change of time could increase the prediction accuracy, but there is still a limitation of the convolution deep neural network model due to spatial smoothing in the strong rainfall region or detailed rainfall prediction.

• The Korean Journal of Applied Statistics
• /
• v.29 no.5
• /
• pp.887-896
• /
• 2016

The Neyman-Scott Rectangular Pulse (NSRP) model is used to model the hourly rainfall series. This model uses a modest number of parameters to represent the rainfall processes and underlying physical phenomena such as the arrival of a storm or rain cells. In this paper, we proposed approximated likelihood function for the NSRP model and applied the proposed method to precipitation data in Seoul.

• Water Engineering Research
• /
• v.1 no.2
• /
• pp.129-136
• /
• 2000

Predicting water levels is a difficult task because a lot of uncertainties are included. Therefore the neural network which is appropriate to such a problem, is introduced. One day ahead forecasting of river stage in the Bocheong River is carried out by using the neural network model. Historical water levels at Snagye gauging point which is located at the downstream of the Bocheong River and average rainfall of the Bocheong River basin are selected as training data sets. With these data sets, the training process has been done by using back propagation algorithm. Then waters levels in 1997 and 1998 are predicted with the trained algorithm. To improve the accuracy, a filtering method is introduced as predicting scheme. It is shown that predicted results are in a good agreement with observed water levels and that a filtering method can overcome the lack of training patterns.