| Home | E-Submission | Sitemap | Contact Us |  
top_img
J. Radiat. Prot. Res > Volume 48(1); 2023 > Article
Lee, Oh, Kim, Jin, and Lee: Development of a Dynamic Downscaling Method for Use in Short-Range Atmospheric Dispersion Modeling Near Nuclear Power Plants

Abstract

Background

High-fidelity meteorological data is a prerequisite for the realistic simulation of atmospheric dispersion of radioactive materials near nuclear power plants (NPPs). However, many meteorological models frequently overestimate near-surface wind speeds, failing to represent local meteorological conditions near NPPs. This study presents a new high-resolution (approximately 1 km) meteorological downscaling method for modeling short-range (<100 km) atmospheric dispersion of accidental NPP plumes.

Materials and Methods

Six considerations from literature reviews have been suggested for a new dynamic downscaling method. The dynamic downscaling method is developed based on the Weather Research and Forecasting (WRF) model version 3.6.1, applying high-resolution land-use and topography data. In addition, a new subgrid-scale topographic drag parameterization has been implemented for a realistic representation of the atmospheric surface-layer momentum transfer. Finally, a year-long simulation for the Kori and Wolsong NPPs, located in southeastern coastal areas, has been made for 2016 and evaluated against operational surface meteorological measurements and the NPPs’ on-site weather stations.

Results and Discussion

The new dynamic downscaling method can represent multiscale atmospheric motions from the synoptic to the boundary-layer scales and produce three-dimensional local meteorological fields near the NPPs with a 1.2 km grid resolution. Comparing the year-long simulation against the measurements showed a salient improvement in simulating near-surface wind fields by reducing the root mean square error of approximately 1 m/s. Furthermore, the improved wind field simulation led to a better agreement in the Eulerian estimate of the local atmospheric dispersion. The new subgrid-scale topographic drag parameterization was essential for improved performance, suggesting the importance of the subgrid-scale momentum interactions in the atmospheric surface layer.

Conclusion

A new dynamic downscaling method has been developed to produce high-resolution local meteorological fields around the Kori and Wolsong NPPs, which can be used in short-range atmospheric dispersion modeling near the NPPs.

Introduction

Nuclear power takes approximately one-third of the low-carbon electricity generation worldwide [1], supplying energy efficiently for households and industries. However, nuclear power plants (NPPs) can significantly impact the public and the environment in case of unexpected nuclear accidents. The radioactive materials released into the atmosphere can cause severe damage near the NPP and downwind regions by short- and long-range transport. For example, the Fukushima Daiichi NPP accident, which occurred after the tsunami in Japan on March 11, 2011, caused significant ongoing radioactive contamination in the vicinity of the site and far downwind area (approximately 80 km) through atmospheric transport [2, 3]. The Japanese authorities took several emergency actions at the early stage of the accident to mitigate the radiological health impact, including a mandatory evacuation of over 200,000 inhabitants near the NPP site and monitoring food, water, and placement of radiation warning signs over the downwind affected areas. In addition, the long-term health and environmental impacts are ongoing over the radioactive contamination areas. For taking timely actions to mitigate the health and environmental impacts under accidental radioactive conditions, it is essential to predict the atmospheric dispersion of the released radioactive materials from real-time meteorological and environmental monitoring data, especially near the accidental site.
Many atmospheric dispersion models with different complexities have been used in representing atmospheric dispersive processes of radioactive materials, such as the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model, the FLEXible PARTicle dispersion model (FLEXPART), and the Gaussian puff model. They were used to analyze the atmospheric dispersion of radioactive materials for the Chernobyl and Fukushima NPP accidents [47]. In addition, the models have also been used to simulate radioactive dispersions of hypothetical accident scenarios for accidental preparedness and NPP site selection [811].
Meanwhile, local meteorology has a crucial influence on atmospheric dispersion; therefore, high-fidelity meteorological data is a prerequisite for reliable atmospheric dispersion modeling of radioactive materials. When radioactive materials are released into the atmosphere, their concentrations can differ depending on the wind velocity, the boundary layer height, and the atmospheric stability [5, 1113]. Lee et al. [7] analyzed the meteorological influences on the long-range transport of the 131I and 137Cs radionuclides using the FLEXPART model and the measurement obtained in the Korean Peninsula during the Fukushima NPP accident. Park et al. [14] analyzed the atmospheric dispersion of 137Cs for hypothetical releases from the Uljin NPP in South Korea using the Lagrangian Particle Dispersion Model and Eulerian Dispersion Model in East Asia, for which they classified synoptic meteorological conditions into 72 classes. The study emphasized the quality of meteorological input data for effective emergency response at the early accidental stage.
In South Korea, NPPs sit on complex terrains close to coastal and mountainous areas. Due to their topographical complexity, the local meteorological fields near the NPPs are formed complicatedly under different synoptic weather conditions [5, 12, 15, 16]. Lee et al. [12] showed that local sea/land breezes were critical in determining the atmospheric concentrations of radioactive materials near the Wolsong NPP. In addition, the importance of wind fields, mainly low-level winds, was emphasized in the atmospheric scalar dispersion, such as radionuclides [5, 15, 16]. However, many meteorological models frequently overestimate wind speeds of the atmospheric boundary layer (ABL), despite alleviation efforts [15], which can result in poor atmospheric dispersion of radioactive materials. In this study, we aim to develop a new high-resolution (approximately 1 km) dynamic downscaling method for modeling short-range (<100 km) atmospheric dispersion of accidental NPP plumes. The Materials and Methods section suggests six considerations for a new dynamic downscaling method from literature reviews and describes the new dynamical downscaling method for the southeastern coastal NPPs in Korea based on the Weather Research and Forecasting (WRF) model with an experimental setup for validation. The Results and Discussion section presents the model’s performance evaluation of the dynamic downscaling method using year-long simulations and observations around the NPPs. The summary and conclusion follow in the last section.

Materials and Methods

1. Considerations for a New Dynamic Downscaling Method

This study targets the Kori and Wolsong NPPs sites for a new dynamic downscaling method located on the southeast coast of the Korean Peninsula. Based on literature reviews, we suggest six considerations for a dynamic downscaling method for modeling short-range (<100 km) atmospheric dispersion from the NPP sites as follows.
First, the meteorological data should be three-dimensional for use in atmospheric dispersion models and have a high spatial resolution to represent complex local circulations near the NPPs. A global climate model (GCM) is adequate for large-scale atmospheric circulations but has limitations due to coarse resolutions (10–100 km) in simulating local meteorological phenomena such as convection, sea/land breezes, and mountain/valley winds. A meteorological downscaling produces the meteorological data with a higher spatio-temporal resolution using the GCM data. Dynamic downscaling methods based on regional atmospheric models are frequently used for three-dimensional, high-resolution local meteorological fields [1719]. They are accomplished by direct numerical integration of a regional atmospheric model using the initial and boundary conditions obtained from the GCM. The dynamical downscaling methods generally require more computing resources than statistical ones; however, they can give high-resolution meteorological data by resolving local circulations [7, 20, 21].
Second, the dynamic downscaling method should sufficiently represent multiscale atmospheric interaction processes from the synoptic scale (approximately 1,000 km) to the ABL scale (approximately 1 km). Many studies have investigated the impact of large-scale atmospheric motions and ABL turbulence on radioactive material dispersions. For example, it has been reported that local circulations associated with synoptic conditions were crucial in determining the atmospheric concentration of radioactive materials [7, 11, 13]. In addition, some studies emphasized the ABL height and the turbulence intensity as sensitive factors in radioactive material dispersions [9, 22]. Since the Kori and Wolsong NPPs are situated in complex terrain areas, they are likely to be affected by local circulations and the ABL under different synoptic weather conditions.
Third, the dynamic downscaling domain should be large enough to cover the area of approximately 100 km from the NPPs so that short-range atmospheric dispersion modeling sufficiently resolves the hypothetical radioactive influences. In the Fukushima Daiichi NPP accident, the radioactive materials significantly contaminated the vicinity of the site and the far downwind area of approximately 80 km [2, 23], issuing a mandatory evacuation within a 50 km area from the NPP [24]. The accident gave us a lesson that the atmospheric dispersion of radioactive materials for emergency response preparedness should be estimated for larger areas from NPPs than in the previous emergency plan.
Forth, a realistic representation of the land use and topography of the dynamic downscaling domain is essential for better simulation of local meteorology. Several studies have found that utilizing realistic land-use data in atmospheric models is beneficial for improving the prediction skill of local meteorology [25, 26]. In addition, other studies showed that the realistic representation of urban areas improved wind speed and direction simulations in the WRF model [27, 28], suggesting the importance of land use information in local wind simulations [29, 30]. However, the default land-use and topography data in the WRF model are based on somewhat old data; thus, static data updates will be necessary.
Fifth, the subgrid-scale surface drag should be adequately represented so that the dynamic downscaling method can represent the fine-scale terrain effects that the model grid cells do not resolve. The turbulent momentum transfer is essential in determining wind fields in the ABL, especially over complex terrain areas. It has been shown that subgrid-scale topographic drag parameterization is an efficient way to simulate better near-surface wind fields in regional atmospheric models [3134]. Jimenez and Dudhia [35] parameterized the subgrid-scale momentum transfer process by adding a forcing term in the momentum equations of the WRF model. Applying the subgrid-scale surface drag over complex terrain areas in the Iberian Peninsula reduced the significant positive bias of near-surface wind speeds found in previous studies [36, 37]. Kim et al. [38] applied it for long-term regional climate simulations over the Korean Peninsula using the WRF model, showing the prediction accuracy of near-surface wind speeds. Lim et al. [15] showed that applying the subgrid-scale orography parameterizations [33, 35] alleviated wind speed biases near the High-flux Advanced Neutron Application Reactor site of the Korea Atomic Energy Research Institute. Despite the improved performance, further reduction of the model’s wind speed biases will be required for better simulation of local meteorology near NPPs.
Sixth, long-term local meteorological data, including meteorological variabilities of annual, seasonal, and hourly timescales, will be needed to estimate reliable short-range (<100 km) atmospheric dispersions near NPPs. In doing so, the dynamic downscaling method requires reliable initial and boundary meteorological conditions from high-fidelity large-scale meteorological forecasts [39, 40].

2. Development of a Dynamical Downscaling Method for the Southeastern Coastal NPPs in Korea

1) A meteorological modeling framework

The dynamic downscaling method is developed based on the WRF model (version 3.6.1) for producing high-resolution local meteorology near the Kori and Wolsong NPPs in a southeastern coastal region in Korea. The WRF model uses fully compressible and non-hydrostatic governing equations in the Arakawa C-grid system. In addition, it includes various physical parameterizations of atmospheric shortwave/longwave radiation transfer, land-atmosphere interaction, ABL turbulence mixing, and grid and subgrid cloud physics [41]. The WRF model, as one of the high-fidelity regional atmospheric models, is suitable for simulating multiscale atmospheric phenomena (e.g., cyclones, typhoons, land/sea breezes). In addition, it can be easily compatible with the atmospheric dispersion models of HYSPLIT and FLEXPART-WRF through model-model coupling interfaces [4244].
The one-way nesting technique is adopted using the WRF model to produce fine-scale local meteorology near the NPPs from large-scale meteorological forecasts. It is a dynamic technique that downscales coarse meteorological data to a fine resolution, and it may be beneficial compared to the two-way (interactive) nesting technique, reducing associated numerical errors [41, 45]. Fig. 1 shows the configuration of four nested grid domains for the Kori and Wolsong NPPs. First, the coarsest domain (denoted as D01) covers a broad East Asia region, including Korea, China, and Japan, to represent synoptic meteorological phenomena on a model grid resolution of 32.4 km. Subsequently, the meteorological fields are repeatedly downscaled to 10.8 and 3.6 km grid resolutions (denoted as D02 and D03), focusing on northeastern Asia and the Korean Peninsula. Finally, the dynamic downscaling method produces high-resolution local meteorology for the NPPs with a grid resolution of 1.2 km (denoted as D04). The finest domain, D04, covers large areas from the NPPs enough to evaluate short-range atmospheric dispersions from the Kori and Wolsong NPPs.
Table 1 summarizes the grid configuration and physical schemes used in the dynamic downscaling method. The model uses 56 sigma vertical levels for a realistic representation of both the synoptic and the ABL meteorological phenomena. The dynamic downscaling method uses physical schemes as follows: the Goddard shortwave radiation scheme [46], the rapid radiative transfer model (RRTM) longwave radiation scheme [47], the Noah land-surface model [48, 49], the Vegetated Urban Canopy Model (VUCM) [50, 51], the Mellor–Yamada–Nakanishi–Niino (MYNN) level 2.5 ABL scheme [5255], the Grell-Devenyi cumulus parameterization [56], and the modified Purdue Lin microphysics scheme [57].
The analysis grid nudging technique is applied using the four-dimensional data assimilation (FDDA) [58] for the D01, D02, and D03 domains, which can reduce forecast errors from the model drift [59]. In this study, the FDDA is applied only for horizontal winds, air temperature, and specific humidity above 10 vertical layers using the ERA (European Centre for Medium-Range Weather Forecasts reanalysis)-Interim reanalysis data from the European Centre for Medium-Range Weather Forecasts. The ERA-Interim reanalysis data is also used for the initial and (lateral) boundary conditions of the D01 domain. It has 6-hourly meteorological fields with a spatial resolution of 0.75°×0.75°.

2) A new subgrid-scale topographic drag parameterization

Reliable simulation of wind fields is essential in estimating short-range atmospheric dispersions from the NPPs. This study added a new subgrid-scale topographic drag parameterization in the WRF model to improve the prediction skill of local wind fields near the NPPs. It accounts for the subgrid-scale topographic variation within the dynamic downscaling grid as an enhanced aerodynamic roughness length (“effective roughness length”). In order to determine the effective roughness length, we used an empirical formulation suggested by Zhu et al. [60]. The empirical formulation was obtained from a series of large-eddy simulations under various hypothetical surface heterogeneities as follows:
(1)
z0σh=α(1+βsh),
where z0 is the aerodynamic roughness length, σh and sh are the standard deviation and the skewness of subgrid topographic heights, and α and β are fitting constants. Here, the constants are set to α=0.1 and β=0.9 from the large-eddy simulation results of Zhu et al. [60]. From the high-resolution topographic height data, σh and sh are calculated as follows:
(2)
σh=[1NxNyΣi=1NxΣj=1Ny(h(xi,yj)-h¯)2]12,
(3)
sh=1NxNyΣi=1NxΣj=1Ny(h(xi,yj)-h¯)3[1NxNyΣi=1NxΣj=1Ny(h(xi,yj)-h¯)2]32.
Here, Nx and Ny are the subgrid numbers of a model grid, h(xi, yj) is the topographic height at a subgrid point (xi, yj), and is the grid mean topographic height. In Results and Discussion section, the performance of the new subgrid-scale topographic drag parameterization will be evaluated by simulating local wind fields near the NPPs.

3) Inclusion of high-resolution land-use and topography datasets

Along with the new subgrid-scale topographic drag parameterization, local land use and topography of the dynamic downscaling domain are updated with high-resolution national datasets. First, we update the land-use data using the 1:25,000 land cover map dataset of the Korea Ministry of Environment (KME). The dataset is made from satellite imagery, aerial photograph, and field survey of the Korean Peninsula, classifying 22 land-use classes. In order to use the KME land-use data in the WRF model, the KME land-use data was converted to be compatible with the United States Geological Survey (USGS) classification of the WRF model. Table 2 shows the mapping table between the KME and the USGS classifications. Then, the converted data were gridded for the WRF Preprocessing System database. The forest regions and the urbanized areas near the NPPs are distinctive from the updated land-use distribution. Meanwhile, we update the model topographic height data using the high-resolution Shuttle Radar Topography Mission (SRTM) topography data (https://www2.jpl.nasa.gov/srtm/) with a spatial resolution of a 3 arc-second (approximately 90 m). Overall, the high-resolution datasets in the dynamic downscaling method represent local land use and topography characteristics near the NPPs and associated physical processes. Fig. 2 shows the topography and land-use distributions of the dynamic downscaling domain using the SRTM topography and the KME land-use dataset.

3. Experimental Setup and Evaluation Method

Long-term local meteorology simulations have been conducted to evaluate the new dynamic downscaling method for the Kori and Wolsong NPPs. In particular, two simulations of the year 2016 were conducted with and without the subgrid-scale topographic drag parameterization (Table 3), and they were compared against the surface measurements obtained from the automatic synoptic observing system (ASOS) stations of the Korea Meteorological Administration (KMA) and the on-site meteorological stations of the NPPs. The analysis focused on investigating the effects of the new ingredient in simulating local winds and local dispersion characteristics around the NPPs. The EXP1 is the simulation without the subgrid-scale topographic drag parameterization, whereas the EXP2 is the simulation with the subgrid-scale topographic drag parameterization. The ASOS sites provide hourly surface measurements, and 10 m wind speed and direction were considered for the model evaluation. Fig. 3 shows the 24 surface measurement sites within the finest dynamic downscaling domain (D04).
The model performance in simulating local winds was evaluated statistically by the mean bias error (MBE), the root mean square error (RMSE), and the Pearson correlation coefficient (R) as follows:
(4)
MBE=1NΣi=1N(Pi-Oi),
(5)
RMSE=1NΣi=1N(Pi-Oi)2,
(6)
R=1σPσOΣi=1N(Pi-P¯)(Oi-O¯),
where Pi and Oi are the ith simulated and measured values, respectively, N is the number of data, and the symbols of overbar (ˉ) and sigma (σ) denote the mean and standard deviation of the data. The MBE value, the average difference between the simulated and measured values, may have positive or negative signs, indicating a better agreement when closer to 0. The RMSE value always has a positive sign, and the lower the value, the better the simulation and measurement agreement. The R measures the linearity strength and direction of the simulated and measured values. R=1 (or R= −1) indicates a strong positive (or negative) correlation, while the value of 0 implies no linear correlation.
Meanwhile, we used the recirculation factor (RF) by Allwine and Whiteman [61] to evaluate the model performance in simulating local dispersion characteristics. The RF is defined using the total distance traveled during a fixed period (S) and the resultant transport distance during the period (L) as follows:
(7)
RF=1-LS.
As shown in Fig. 4, the RF is a time-integral quantity calculated from wind data collected at fixed time intervals at single meteorological stations [61, 62]. This study estimated the RF quantities every day using hourly simulated and measured near-surface wind fields at every measurement station, and the quantity ranges from 0 to 1. As RF is close to 1, the endpoint of an air parcel returns to the starting point (high recirculation) (Fig. 4A). In contrast, when RF is close to 0, the air parcel transports with little recirculation (Fig. 4B). In Allwine and Whiteman [61], local flows with daily RF >0.6 indicate that recirculation dominates, whereas local flows with daily RF <0.2 indicate that ventilation dominates. Overall, the quantity can give useful local dispersion characteristics at individual measurement stations under the assumption of homogeneous local wind conditions.

Results and Discussion

1. Evaluation of Local Winds

The new dynamic downscaling method produced year-long three-dimensional local meteorological fields near the Kori and Wolsong NPPs with a 1.2 km grid resolution, representing multiscale atmospheric motions from the synoptic to the boundary-layer scales. Fig. 5 shows statistical evaluation results of the simulated near-surface wind speeds against the 22 ASOS measurements. The simulated values at the nearest grid point of the measurement stations were used for the model-measurement comparison, and the MBE and RMSE were calculated for the two experiments with or without the new subgrid-scale topographic drag parameterization. The comparison was made for all-day, daytime (0900–1500 local standard time [LST]), and nighttime (2100–next day 0300 LST) periods. The simulation without the subgrid-scale topographic drag parameterization (EXP1) overestimated the measured wind speeds for most stations by approximately 1 m/s on average (Fig. 5A). This is due to the absence of surface drag by unresolved (subgrid-scale) topographic height variations. The overestimation of wind speed was more significant in the winter than in the summertime. The RMSE values range from 1.6 m/s in June to 2.2 m/s in January (Fig. 5B). In addition, the model overestimated wind speeds in the daytime than in the nighttime (Fig. 5C–5F). The higher biases in wintertime and the daytime are attributed to relatively higher wind speeds in wintertime and the daytime than in summertime and the nighttime, respectively. This result is consistent with previous studies that reported overestimated near-surface wind speed in the WRF model [3335, 38]. Meanwhile, the simulation with the subgrid-scale topographic drag parameterization (EXP2) reduced the model’s high wind speed biases significantly, ranging the MBE values within ±0.3 m/s on average (Fig. 5A). The RMSE values range from 1.1 m/s in November to 1.4 m/s in April (Fig. 5B). In addition, the improvement of the model performance in simulating local winds was apparent in both the daytime and nighttime (Fig. 5C–5F).
The model performance in simulating local winds was further evaluated for the u-wind (eastward) and v-wind (northward) components measured at the 22 ASOS stations. The subgrid-scale topographic drag parameterization is applied to both u- and v-wind components of modeled winds, so the parameterization effect differs in the independent wind components. First, Fig. 6 shows the MBE and RMSE values for the u-wind components simulated with and without the subgrid-scale topographic drag parameterization. The simulation EXP1 showed good agreement with the measurements in the u-wind component (Fig. 6A), but it slightly overestimated the measurements in wintertime (January, February, December) and underestimated in autumn (September, October). The mean RMSE values range from 1.5 m/s in July to 2.3 m/s in February (Fig. 6B). Furthermore, the model biases appeared more apparently in the daytime than in the nighttime each month (Fig. 6C, 6E). On the other hand, the simulation EXP2 showed better agreement with the measurements, alleviating the model errors significantly (Fig. 6A). Primarily, it reduced positive biases in the wintertime and negative biases in the autumn while maintaining good performance in the other seasons. The RMSE values range 1.1–1.5 m/s, lower by approximately 0.4–0.8 m/s than the simulation EXP1 (Fig. 6B). In addition, the improvement of the model performance was apparent in both the daytime (Fig. 6C, 6D) and the nighttime (Fig. 6E, 6F). Similarly, Fig. 7 shows the MBE and RMSE values for the v-wind components simulated with and without the subgrid-scale topographic drag parameterization. The simulation EXP1 showed slightly negative biases except for April to July, indicating that the model tends to simulate the northerly winds stronger than the measurements, especially in wintertime (Fig. 7A). The error may attribute to the seasonal variations of synoptic conditions in the Korean Peninsula [63]. In addition, the model biases persisted for both daytime and nighttime each month (Fig. 7C, 7E). The mean RMSE values range from 1.6 m/s in June to 2.2 m/s in April (Fig. 7B). In contrast, the simulation EXP2 showed apparent model performance improvements, as shown similarly in Figs. 5, 6, emphasizing the importance of the subgrid-scale topographic drag parameterization over complex terrain areas (Fig. 7B, 7D, 7F).
We further evaluated the model performance in simulating local winds by comparing it against the on-site measurements at the Kori and Wolsong NPPs. Fig. 8 shows the RMSE values for the near-surface winds of 2016 simulated with and without the subgrid-scale topographic drag parameterization. In addition, Table 4 presents the statistical evaluation results at the Kori and Wolsong NPP sites for each month in 2016. In the simulation EXP1, the dynamic downscaling method performed well for the Kori and Wolsong NPP sites, but the model overpredicted the near-surface wind speed at the Wolsong NPP site in wintertime, leading to a maximum RMSE of 4.3 m/s in January. As a result, the annual mean RMSE values were 1.9 m/s in Kori and 2.6 m/s at the Wolsong site. Meanwhile, the simulation EXP2 showed better agreement than EXP1, with the RMSE values of 1.4 m/s in Kori and 1.3 m/s at the Wolsong site. In addition, the simulation EXP2 significantly improved the poor wintertime model performance of EXP1, reducing the RMSE from 4.3 to 1.4 m/s at the Wolsong site in January. These results confirm that the subgrid-scale topographic drag process improves the model performance in simulating local winds near the NPP sites, showing the potential of the new dynamic downscale model for local wind simulations.

2. Evaluation of Local Atmospheric Dispersion Characteristics

Based on the improved model performance in simulating local winds, we investigated the impact of the high-resolution local winds on short-range atmospheric dispersion near the Kori and Wolsong NPP sites. The daily RF quantities were calculated for the year of 2016, using hourly simulated and measured near-surface wind fields at the surface measurement stations. Fig. 9 compares the spatial distributions of the annual mean RF measured and simulated with and without the subgrid-scale topographic drag parameterization. The RF values calculated from the surface measurements ranged from 0.3 to 0.6 over the Kori and Wolsong NPPs domain (Fig. 9A). The coastal areas near the Kori and Wolsong NPPs had relatively low RF values of approximately 0.2, while recirculation dominates inland mountainous areas. In simulation EXP1, the RF values were lower than 0.3 for most domain areas and less than 0.2 at the Kori and Wolsong NPP sites (Fig. 9B). Meanwhile, the RF values in the simulation EXP2 increased up to 0.3–0.6 over inland mountainous areas, reducing the model-measurement discrepancy in the EXP1. Fig. 10 shows the scatter distribution of the measured and simulated annual-mean RF values estimated at the meteorological stations. As shown in Fig. 9, the simulation EXP2 shows a better performance than the EXP1. Applying the subgrid-scale topographic drag parameterization enhanced the model performance in simulating short-range atmospheric dispersion characteristics, increasing the correlation coefficient from 0.56 in the EXP1 to 0.71 in the EXP2.

Conclusion

An accurate assessment of the atmospheric dispersion of radioactive materials is necessary to establish emergency response plans for the vicinity of NPPs. In this study, we developed a new high-resolution (approximately 1 km) dynamic downscaling method for modeling short-range (<100 km) atmospheric dispersion from the Kori and Wolsong NPPs in Korea. Six considerations from literature reviews have been reflected on the dynamic downscaling method for high-fidelity meteorological data. The dynamic downscaling method is developed based on the WRF model version 3.6.1, applying high-resolution KME land-use and SRTM topography data for the Kori and Wolsong NPPs domain. In addition, a new subgrid-scale topographic drag parameterization has been implemented for a realistic representation of the atmospheric surface-layer momentum transfer process. Finally, using year-long simulations with and without the subgrid-scale topographic drag parameterization, the model performance in simulating local winds and atmospheric dispersion characteristics has been evaluated against the KMA operational surface meteorological data and the NPP on-site measurements.
The model validation results showed that the new dynamic downscaling method could produce reliable local winds and atmospheric dispersion characteristics near the Kori and Wolsong NPPs mainly by applying the new subgrid-scale topographic drag parameterization. Meteorological models frequently overestimate near-surface wind speeds over complex terrain areas [3335]. This study found that updating land-use and topography data, as in many previous studies [2530], has a limitation in reducing common model biases over complex terrain without considering the subgrid-scale topographic drag parameterization. It suggests that subgrid-scale turbulent momentum interaction is essential in modeling high-resolution local winds over complex terrain areas. Because of the good meteorological performance of the dynamic downscaling method, the model reasonably reproduced the atmospheric dispersion characteristics estimated from the near-surface wind measurements over the domain.
Overall, the new dynamic downscaling method showed the potential to produce high-resolution local meteorological fields over the Kori and Wolsong NPPs domain. Furthermore, it can generally be applied to producing high-fidelity meteorological data over complex terrain areas for atmospheric dispersion modeling purposes. The short-range (<100 km) atmospheric dispersion characteristics of radioactive materials from the Kori and Wolsong NPPs will be investigated using atmospheric dispersion models (e.g., HYSPLIT, FLEXPART).

Acknowledgements

This research was supported by the Nuclear Safety Research Program through the Korea Foundation Of Nuclear Safety (KoFONS) funded by the Nuclear Safety and Security Commission (No. 1805018).

Notes

Conflict of Interest

No potential conflict of interest relevant of BomIn Science Consulting to this article was reported.

Ethical Statement

This article does not contain any studies with human participants or animals performed by any of the authors.

Author Contributions

Conceptualization: Lee SH, Kim CJ. Methodology: Lee SH. Formal analysis: Lee SH, Jin CS. Funding acquisition: Lee HH. Writing - original draft: Lee SH, Oh SB. Writing - review and editing: Jin CS, Oh SB. Approval of final manuscript: all authors.

References

1. Schneider M, Froggatt A. World nuclear industry status report 2020 [Internet]. Paris, France, A Mycle Schneider Consulting Project; 2020 [cited 2022 Dec 19] Available from: https://www.worldnuclearreport.org/-World-Nuclear-Industry-Status-Report-2020-.html

2. Dauer LT, Zanzonico P, Tuttle RM, Quinn DM, Strauss HW. The Japanese tsunami and resulting nuclear emergency at the Fukushima Daiichi power facility: technical, radiologic, and response perspectives. J Nucl Med. 2011;52(9):1423-1432.
crossref pmid
3. United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation (Scientific Annex B:). New York, NY, United Nations Scientific Committee on the Effects of Atomic Radiation; 2000 [cited 2022 Dec 19]. Available from: https://www.unscear.org/docs/publications/2020/UNSCEAR_2020_AnnexB_AdvanceCopy.pdf

4. Stohl A, Seibert P, Wotawa G, Arnold D, Burkhart JF, Eckhardt S, et al. Xenon-133 and caesium-137 releases into the atmosphere from the Fukushima Dai-ichi nuclear power plant: determination of the source term, atmospheric dispersion, and deposition. Atmos Chem Phys. 2012;12(5):2313-2343.
crossref
5. Srinivas CV, Venkatesan R, Baskaran R, Rajagopal V, Venkatraman B. Regional scale atmospheric dispersion simulation of accidental releases of radionuclides from Fukushima Dai-ichi reactor. Atmos Environ. 2012;61:66-84.
crossref
6. Korsakissok I, Mathieu A, Didier D. Atmospheric dispersion and ground deposition induced by the Fukushima Nuclear Power Plant accident: a local-scale simulation and sensitivity study. Atmos Environ. 2013;70:267-279.
crossref
7. Lee HJ, Jo HY, Nam KP, Lee KH, Kim CH. Measurement, simulation, and meteorological interpretation of medium-range transport of radionuclides to Korea during the Fukushima Dai-ichi nuclear accident. Ann Nucl Energy. 2017;103:412-423.
crossref
8. Thaning L, Baklanov A. Simulation of the atmospheric transport and deposition on a local/meso-and regional scale after hypothetical accidents at the Kola nuclear power plant. Sci Total Environ. 1997;202(1–3):199-210.
crossref
9. Suh KS, Kim EH, Han MH. Development of long-range atmospheric dispersion model against a nuclear accident. J Radiat Prot Res. 2002;27(3):171-179.

10. Srinivas CV, Venkatesan R. A simulation study of dispersion of air borne radionuclides from a nuclear power plant under a hypothetical accidental scenario at a tropical coastal site. Atmos Environ. 2005;39(8):1497-1511.
crossref
11. Kim CH, Song CK, Lee SH, Song SK. Simulating mesoscale transport and diffusion of radioactive noble gases using the Lagrangian particle dispersion model. J Environ Radioact. 2008;99(10):1644-1652.
crossref pmid
12. Lee GB, Lee MC, Song YI. A study on mesoscale atmospheric dispersion of radioactive particles released from nuclear power plants. J Radiat Prot Res. 1997;22(4):273-288.

13. Jeong H, Kim E, Park M, Jeong H, Hwang W, Han M. Numerical simulation of air pollutant dispersion using an in situ tracer experiment at a nuclear site. Ann Nucl Energy. 2014;73:1-6.
crossref
14. Park SU, Lee IH, Joo SJ, Ju JW. Emergency preparedness for the accidental release of radionuclides from the Uljin Nuclear Power Plant in Korea. J Environ Radioact. 2017;180:90-105.
crossref
15. Lim KS, Lim JM, Shin HH, Hong J, Ji YY, Lee W. Impacts of subgrid-scale orography parameterization on simulated atmospheric fields over Korea using a high-resolution atmospheric forecast model. Meteorol Atmos Phys. 2019;131(4):975-985.
crossref
16. Hanna SR. The exponential probability density function and concentration fluctuations in smoke plumes. Bound Layer Meteorol. 1984;29(4):361-375.
crossref
17. Giorgi F, Marinucci MR, Visconti G. Use of a limited-area model nested in a general circulation model for regional climate simulation over Europe. J Geophys Res Atmos. 1990;95(D11):18413-18431.
crossref
18. Gibelin AL, Deque M. Anthropogenic climate change over the Mediterranean region simulated by a global variable resolution model. Clim Dyn. 2003;20(4):327-339.
crossref
19. Boe J, Terray L, Habets F, Martin E. Statistical and dynamical downscaling of the Seine basin climate for hydro-meteorological studies. Int J Climatol. 2007;27(12):1643-1655.
crossref
20. Lo JC, Yang ZL, Pielke RA Sr. Assessment of three dynamical climate downscaling methods using the Weather Research and Forecasting (WRF) model. J Geophys Res Atmos. 2008;113(D9):D09112.
crossref
21. Ahn JB, Lee J, Im ES. The reproducibility of surface air temperature over South Korea using dynamical downscaling and statistical correction. J Meteorol Soc Jpn. 2012;90(4):493-507.
crossref
22. Hwang WT, Kim EH, Jeong HS, Jeong HJ, Han MH. Influence of statistical compilation of meteorological data on short-term atmospheric dispersion factors in a hypothetical accidental release of nuclear power plants. J Radiat Prot Res. 2012;37(3):116-122.
crossref
23. United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation (UNSCEAR 2008). New York, NY, United Nations Scientific Committee on the Effects of Atomic Radiation; 2011 [cited 2022 Dec 19]. Available from: https://www.unscear.org/docs/reports/2008/11-80076_Report_2008_Annex_D.pdf

24. Kim TY, Lee KH, Chae S, Ha SH, Chung SW, Lee UJ. Development of training system for radiation exposure reduction and protection of local residents against radioactivity release accidents at nuclear power plants. J Korean Soc Hazard Mitig. 2019;19(2):79-90.
crossref
25. Civerolo K, Hogrefe C, Lynn B, Rosenthal J, Ku JY, Solecki W, et al. Estimating the effects of increased urbanization on surface meteorology and ozone concentrations in the New York City metropolitan region. Atmos Environ. 2007;41(9):1803-1818.
crossref
26. Cheng FY, Byun DW. Application of high resolution land use and land cover data for atmospheric modeling in the Houston–Galveston metropolitan area, Part I: meteorological simulation results. Atmos Environ. 2008;42(33):7795-7811.
crossref
27. De Meij AA, Vinuesa JF. Impact of SRTM and Corine Land Cover data on meteorological parameters using WRF. Atmos Res. 2014;143:351-370.
crossref
28. Santos-Alamillos FJ, Pozo-Vazquez D, Ruiz-Arias JA, Tovar-Pescador J. Influence of land-use misrepresentation on the accuracy of WRF wind estimates: evaluation of GLCC and CORINE land-use maps in southern Spain. Atmos Res. 2015;157:17-28.
crossref
29. Cheng FY, Hsu YC, Lin PL, Lin TH. Investigation of the effects of different land use and land cover patterns on mesoscale meteorological simulations in the Taiwan area. J Appl Meteorol Climatol. 2013;52(3):570-587.
crossref
30. Cuxart J, Wrenger B, Martinez-Villagrasa D, Reuder J, Jonassen MO, Jimenez MA, et al. Estimation of the advection effects induced by surface heterogeneities in the surface energy budget. Atmos Chem Phys. 2016;16(14):9489-9504.
crossref
31. Georgelin M, Bougeault P, Black T, Brzovic N, Buzzi A, Calvo J, et al. The second COMPARE exercise: a model intercomparison using a case of a typical mesoscale orographic flow, the PYREX IOP3. Q J R Meteorol Soc. 2000;126(564):991-1029.
crossref
32. Rontu L. A study on parametrization of orography-related momentum fluxes in a synoptic-scale NWP model. Tellus A Dyn Meteorol Oceanogr. 2006;58(1):69-81.
crossref
33. Mass C, Ovens D. WRF model physics: problems, solutions and a new paradigm for progress [Internet]. Boulder, CO, National Center for Atmospheric Research; 2010 [cited 2022 May 9]. Available from: https://www.atmos.washington.edu/~cliff/WRFworkshop2010b.ppt

34. Lee J, Shin HH, Hong SY, Jimenez PA, Dudhia J, Hong J. Impacts of subgrid-scale orography parameterization on simulated surface layer wind and monsoonal precipitation in the high-resolution WRF model. J Geophys Res Atmos. 2015;120(2):644-653.
crossref
35. Jimenez PA, Dudhia J. Improving the representation of resolved and unresolved topographic effects on surface wind in the WRF model. J Appl Meteorol Climatol. 2012;51(2):300-316.
crossref
36. Cheng WY, Steenburgh WJ. Evaluation of surface sensible weather forecasts by the WRF and the Eta models over the western United States. Weather Forecast. 2005;20(5):812-821.
crossref
37. Bernardet L, Nance L, Demirtas M, Koch S, Szoke E, Fowler T, et al. The Developmental Testbed Center and its winter forecasting experiment. Bull Am Meteorol Soc. 2008;89(5):611-628.
crossref
38. Kim JY, Kim DY, Oh JH, Kim SH, Kim HG, Kang YH, et al. Sensitivity evaluation of surface wind simulations by surface drag parameterization and spatial resolution using WRF model. J Wind Eng Inst Korea. 2015;19(3):77-83.

39. Diro GT, Tompkins AM, Bi X. Dynamical downscaling of ECMWF ensemble seasonal forecasts over East Africa with RegCM3. J Geophys Res Atmos. 2012;117(D16):D16103.
crossref
40. Uppala S, Dee D, Kobayashi S, Berrisford P, Simmons A. Towards a climate data assimilation system: status update of ERA-Interim. ECMWF Newsl. 2008;115(7):12-18.

41. Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker DM, Wang W, et al. A description of the Advanced Research WRF version 3 [Internet]. Boulder, CO, National Center for Atmospheric Research; 2008 [cited 2022 May 9]. Available from: https://doi.org/10.5065/D68S4MVH

42. Draxler RR, Hess GD. Description of the HYSPLIT_4 modeling system [Internet]. Silver Spring, MD, Air Resources Laboratory; 1997 [cited 2022 Dec 19]. Available from: https://www.arl.noaa.gov/documents/reports/arl-224.pdf

43. Draxler RR, Hess GD. An overview of the HYSPLIT_4 modelling system for trajectories, dispersion, and deposition. Aust Meteorol Mag. 1998;47(4):295-308.

44. Brioude J, Arnold D, Stohl A, Cassiani M, Morton D, Seibert P, et al. The Lagrangian particle dispersion model FLEXPART-WRF version 3.1. Geosci Model Dev. 2013;6(6):1889-1904.
crossref
45. Sundstrom A, Elvius T. Computational problems related to limited area modelling. In: Mesinger F, Arakawa A. Numerical methods used in atmospheric models. Geneva, Switzerland, World Meteorological Organization, International Council of Scientific Unions. 1976;379-416.

46. Chou MD, Suarez MJ, Ho CH, Yan MM, Lee KT. Parameterizations for cloud overlapping and shortwave single-scattering properties for use in general circulation and cloud ensemble models. J Clim. 1998;11(2):202-214.
crossref
47. Mlawer EJ, Taubman SJ, Brown PD, Iacono MJ, Clough SA. Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J Geophys Res Atmos. 1997;102(D14):16663-16682.
crossref
48. Chen F, Dudhia J. Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part II: preliminary model validation. Mon Weather Rev. 2001;129(4):587-604.
crossref
49. Ek MB, Mitchell KE, Lin Y, Rogers E, Grunmann P, Koren V, et al. Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J Geophys Res Atmos. 2003;108(D22):8851.
crossref
50. Lee SH, Park SU. A vegetated urban canopy model for meteorological and environmental modelling. Bound Layer Meteorol. 2008;126(1):73-102.
crossref
51. Lee SH, Lee H, Park SB, Woo JW, Lee DI, Baik JJ. Impacts of in-canyon vegetation and canyon aspect ratio on the thermal environment of street canyons: numerical investigation using a coupled WRF-VUCM model. Q J R Meteorol Soc. 2016;142(699):2562-2578.
crossref
52. Nakanish M. Improvement of the Mellor–Yamada turbulence closure model based on large-eddy simulation data. Bound Layer Meteorol. 2001;99(3):349-378.
crossref
53. Nakanishi M, Niino H. An improved Mellor–Yamada level-3 model with condensation physics: its design and verification. Bound Layer Meteorol. 2004;112(1):1-31.
crossref
54. Nakanishi M, Niino H. An improved Mellor–Yamada level-3 model: its numerical stability and application to a regional prediction of advection fog. Bound Layer Meteorol. 2006;119(2):397-407.
crossref
55. Nakanishi M, Niino H. Development of an improved turbulence closure model for the atmospheric boundary layer. J Meteorol Soc Jpn. 2009;87(5):895-912.
crossref
56. Grell GA, Devenyi D. A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys Res Lett. 2002;29(14):1693.
crossref
57. Lin YL, Farley RD, Orville HD. Bulk parameterization of the snow field in a cloud model. J Appl Meteorol Climatol. 1983;22(6):1065-1092.
crossref
58. Stauffer DR, Seaman NL. Use of four-dimensional data assimilation in a limited-area mesoscale model. Part I: experiments with synoptic-scale data. Mon Weather Rev. 1990;118(6):1250-1277.
crossref
59. Mun J, Lee HW, Jeon W, Lee SH. Impact of meteorological initial input data on WRF simulation: comparison of ERA-interim and FNL data. J Environ Sci Int. 2017;26(12):1307-1319.
crossref
60. Zhu X, Iungo GV, Leonardi S, Anderson W. Parametric study of urban-like topographic statistical moments relevant to a priori modelling of bulk aerodynamic parameters. Bound Layer Meteorol. 2017;162(2):231-253.
crossref
61. Allwine KJ, Whiteman CD. Single-station integral measures of atmospheric stagnation, recirculation and ventilation. Atmos Environ. 1994;28(4):713-721.
crossref
62. Halios CH, Flocas HA, Helmis CG, Asimakopoulos DN, Mouschouras PG. Observations of local meteorological variability under large-scale circulation patterns over Athens, Greece. Atmosphere. 2018;9(1):25.
crossref
63. Oh JS, Lee JH, Woo JW, Lee DI, Lee SH, Seo J, et al. Performance evaluation of the high-resolution WRF meteorological simulation over the Seoul metropolitan area. Atmosphere. 2020;30(3):257-276.

Fig. 1
Domain configuration of the dynamic downscaling method for the Kori and Wolsong nuclear power plants. Shading denotes the topographic height.
jrpr-2022-00073f1.jpg
Fig. 2
Topography and land-use distribution of the dynamic downscaling domain 4 (D04) updated by the Korea Ministry of Environment land-use dataset. Circular symbols denote the Kori and Wolsong nuclear power plant sites.
jrpr-2022-00073f2.jpg
Fig. 3
Location of the automatic synoptic observing system (ASOS) stations (red circles) and the on-site meteorological stations of the Kori and Wolsong nuclear power plants (yellow triangles with a black dot) within the finest dynamic downscaling domain (D04).
jrpr-2022-00073f3.jpg
Fig. 4
Illustration of the high-recirculation (A) and low-recirculation (B) conditions estimated by the total distance traveled during a fixed period “S” and the resultant transport distance during the period “L”. A four-digit number denotes the local standard time (LST) with a time interval of 1-hour during a single day. RF, recirculation factor.
jrpr-2022-00073f4.jpg
Fig. 5
Statistical evaluation results of near-surface wind speeds simulated with and without the subgrid-scale topographic drag parameterization in terms of MBE (A, C, E) and RMSE (B, D, E). The daytime period is 0900–1500 LST, and the nighttime is 2100–0300 LST. MBE, mean bias error; RMSE, root mean square error; LST, local standard time.
jrpr-2022-00073f5.jpg
Fig. 6
Statistical evaluation results of near-surface u-wind simulated with and without the subgrid-scale topographic drag parameterization in terms of MBE (A, C, E) and RMSE (B, D, E). The daytime period is 0900–1500 LST, and the nighttime is 2100–0300 LST. MBE, mean bias error; RMSE, root mean square error; LST, local standard time.
jrpr-2022-00073f6.jpg
Fig. 7
Statistical evaluation results of near-surface v-wind simulated with and without the subgrid-scale topographic drag parameterization in terms of MBE (A, C, E) and RMSE (B, D, E). The daytime period is 0900–1500 LST, and the nighttime is 2100–0300 LST. MBE, mean bias error; RMSE, root mean square error; LST, local standard time.
jrpr-2022-00073f7.jpg
Fig. 8
Statistical evaluation results of 10 m wind speed of all-component, u-wind, and v-wind in RMSE using meteorological measurement station of (A) Kori and (B) Wolsong nuclear power plants (NPPs). RMSE, root mean square error.
jrpr-2022-00073f8.jpg
Fig. 9
Spatial distributions of the annual mean recirculation factor values estimated from (A) the surface measurements, (B) the simulation EXP1, and (C) the simulation EXP2.
jrpr-2022-00073f9.jpg
Fig. 10
Comparison of the annual mean recirculation factor values estimated from the measurement and the simulations EXP1 and EXP2.
jrpr-2022-00073f10.jpg
Table 1
Model Grid Configuration and Physical Schemes for the Dynamic Downscaling Method
D01 D02 D03 D04
Horizontal grid (grid spacing) 181×134 (32.4 km) 190×181 (10.8 km) 181×190 (3.6 km) 181×217 (1.2 km)
Vertical grid 56 levels
Shortwave radiation Goddard [46]
Longwave radiation RRTM [47]
Land surface process Noah LSM [48, 49]
Urban canopy model VUCM [50, 51]
Turbulence process MYNN Level 2.5 [5255]
Cumulus parameterization Grell-Devenyi [56]
Cloud physics Modified Purdue Lin [57]
Four-dimensional data assimilation (FDDA) O O O X

RRTM, rapid radiative transfer model; Noah LSM, Noah land-surface model; VUCM, vegetated urban canopy model; MYNN, Mellor–Yamada–Nakanishi–Niino.

Table 2
Mapping Table between KME Land-Use Data and USGS Land-Use Data
KME land-use classification USGS land-use classification
Residential area (110) Low-intensity residential (31)
Industrial area (120) Industrial/Commercial (33)
Commercial area (130) Industrial/Commercial (33)
Entertainment facility area (140) Industrial/Commercial (33)
Traffic area (150) Industrial/Commercial (33)
Public facility (160) Low-intensity residential (31)
Rice pad (210) Irrigated cropland and pasture (3)
Field (220) Dryland cropland and pasture (2)
Vinyl house cultivation area (230) Dryland cropland and pasture (2)
Orchard (240) Dryland cropland and pasture (2)
Other cultivation areas (250) Dryland cropland and pasture (2)
Broadleaf tree forest (310) Evergreen broadleaf forest (13)
Needleleaf tree forest (320) Evergreen needleleaf forest (14)
Mixed tree forest (330) Mixed forest (15)
Natural green land (410) Grassland (7)
Golf course (420) Grassland (7)
Inland wetland (510) Herbaceous wetland (17)
Coastal wetland (520) Herbaceous wetland (17)
Mining area (610) Bare ground tundra (23)
Other bare lands (620) Bare ground tundra (23)
Inland water (710) Lakes (28)
Seawater (720) Water bodies (16)

The number in parentheses denotes the classification code.

KME, Korea Ministry of Environment; USGS, United States Geological Survey.

Table 3
Two Sensitivity Simulations of the Dynamic Downscaling Method
EXP1 EXP2
Integration period January 1–December 31, 2016 January 1–December 31, 2016
Subgrid-scale topographic drag parameterization X O
Table 4
Statistical Evaluation Results of 10 m Wind Speed of All-Component, U-Wind, and V-Wind in RMSE (m/s) for Each Month in 2016 using Meteorological Measurement Station of Kori and Wolsong NPPs
Kori NPP Wolsong NPP


Wind speed U-wind V-wind Wind speed U-wind V-wind






EXP1 EXP2 EXP1 EXP2 EXP1 EXP2 EXP1 EXP2 EXP1 EXP2 EXP1 EXP2
Jan 1.6 1.7 1.8 1.5 1.7 1.5 4.3 1.4 4.2 1.8 2.4 1.1

Feb 2.0 1.7 1.9 1.5 2.2 1.8 3.5 1.4 3.4 1.7 2.4 1.2

Mar 1.8 1.8 1.6 1.3 2.2 1.9 2.2 1.3 2.4 1.7 2.1 1.3

Apr 2.1 1.5 1.8 1.3 2.3 1.6 2.4 1.2 2.2 1.7 2.3 1.2

May 1.8 1.6 1.5 1.4 2.2 1.6 1.8 1.3 1.7 1.5 2.0 1.2

Jun 1.8 1.2 1.6 1.1 2.0 1.4 1.7 1.1 1.7 1.3 1.7 1.1

Jul 1.9 1.4 1.4 1.1 2.2 1.6 1.6 1.0 1.5 1.2 2.0 1.3

Aug 1.6 1.2 1.2 1.0 1.9 1.2 1.8 1.1 1.7 1.4 1.7 0.9

Sep 2.1 1.1 2.0 1.1 2.0 1.3 2.0 1.3 1.7 1.5 2.2 1.3

Oct 2.1 1.1 1.8 1.1 2.1 1.4 2.1 1.3 1.9 1.6 2.0 1.2

Nov 1.8 1.4 1.6 1.1 2.0 1.6 2.9 1.5 2.8 1.7 2.2 1.2

Dec 1.6 1.4 1.6 1.3 1.8 1.4 3.0 1.3 2.8 1.5 2.3 1.2

Mean 1.9 1.4 1.6 1.2 2.0 1.5 2.4 1.3 2.3 1.6 2.1 1.2

RMSE, root mean square error; NPP, nuclear power plant.

Editorial Office
#319, Hanyang Institute of Technology Bldg., 222 Wangsimni-ro, Seongdong-gu,Seoul, Republic of Korea
Tel: +82-2-2297-9775   Fax: +82-2-2297-9776
Email: managing.editor@jrpr.org
About |  Browse Articles |  Current Issue |  For Authors and Reviewers
Copyright © by Korean Association for Radiation Protection. Developed in M2PI