DOI QR코드

DOI QR Code

A Case Study: Improvement of Wind Risk Prediction by Reclassifying the Detection Results

풍해 예측 결과 재분류를 통한 위험 감지확률의 개선 연구

  • 김수옥 (에스티에이코퍼레이션(주)) ;
  • 황규홍 (에스티에이코퍼레이션(주))
  • Received : 2021.04.08
  • Accepted : 2021.09.30
  • Published : 2021.09.30

Abstract

Early warning systems for weather risk management in the agricultural sector have been developed to predict potential wind damage to crops. These systems take into account the daily maximum wind speed to determine the critical wind speed that causes fruit drops and provide the weather risk information to farmers. In an effort to increase the accuracy of wind risk predictions, an artificial neural network for binary classification was implemented. In the present study, the daily wind speed and other weather data, which were measured at weather stations at sites of interest in Jeollabuk-do and Jeollanam-do as well as Gyeongsangbuk- do and part of Gyeongsangnam- do provinces in 2019, were used for training the neural network. These weather stations include 210 synoptic and automated weather stations operated by the Korean Meteorological Administration (KMA). The wind speed data collected at the same locations between January 1 and December 12, 2020 were used to validate the neural network model. The data collected from December 13, 2020 to February 18, 2021 were used to evaluate the wind risk prediction performance before and after the use of the artificial neural network. The critical wind speed of damage risk was determined to be 11 m/s, which is the wind speed reported to cause fruit drops and damages. Furthermore, the maximum wind speeds were expressed using Weibull distribution probability density function for warning of wind damage. It was found that the accuracy of wind damage risk prediction was improved from 65.36% to 93.62% after re-classification using the artificial neural network. Nevertheless, the error rate also increased from 13.46% to 37.64%, as well. It is likely that the machine learning approach used in the present study would benefit case studies where no prediction by risk warning systems becomes a relatively serious issue.

농업기상재해 조기경보시스템에서는 일 최대순간 풍속에 과수의 낙과 피해 임계풍속을 대입하여 농작물의 풍해 위험을 예측, 자원농가에게 제공하고 있다. 강풍의 위험 예측확률을 높이기 위한 방법으로써, 기존 방식에서 '안전'으로 분류된 데이터들 중 실제로는 풍해위험이 있는 경우를 찾아내는 인공신경망 이항분류 기법을 도입하였다. 학습데이터는 전라남북도와 경북 및 경남 일부지역의 총 210개소 기상청 종관 및 방재기상관측지점에서 수집된 2019년 전체 일별 풍속자료이며, 최적 모델 도출을 위한 검증데이터는 동일지점의 2020년 1월 1일~12월 12일 자료를, 인공신경망 기법 사용 전/후의 풍해위험예측 성능 평가는 2020년 12월 13일~2021년 2월 18일까지의 자료를 사용하였다. 풍해위험 임계풍속은 과수의 낙과 피해기준으로 주로 사용되고 있는 11m/s를 설정하였다. 또한 2020년 동일 기간의 일 최대순간풍속 실측값으로 Weibull 분포를 작성한 후, 추정값과 임계풍속간의 편차를 이용하여 누적확률값을 계산, 풍해 경보에서 한 단계 낮은 주의보를 판단하고 인공신경망 기법 적용 결과와 비교하였다. 평가기간 중 기존의 풍해 위험 탐지확률은 65.36%였으나 인공신경망 기법으로 재분류 과정을 거친 후 93.62%로 크게 개선되었다. 반면, 오보율이 함께 증가되어(13.46% → 37.64%), 전반적인 정확도는 감소하였다. 한편 Weibull 분포를 이용하여 풍해주의보 구간을 두었을 때는 정확도 83.46%으로 인공신경망 기법에 비해 전반적인 예측 정확도는 더 높았던 반면 위험 탐지확률은 88.79%로 더 낮게 나타났다. 따라서, 상대적으로 위험예보의 미예측이 중대한 문제가 되는 사례에서 인공신경망 방식이 유용할 것으로 보인다.

Keywords

Acknowledgement

본 논문은 농촌진흥청 연구사업 신농업기후변화대응체계구축 (과제번호: PJ01500703)의 지원에 의해 이루어진 것임.

References

  1. Kim, S. O., and J. I. Yun, 2011: A quantification method for the cold pool effect on nocturnal temperature in a closed catchment. Korean Journal of Agricultural and Forest Meteorology 13(4), 176-184. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2011.13.4.176
  2. Kim, S. O., J. H. Kim, D. J. Kim, and J. I. Yun, 2012: Wind effect on daily minimum temperature across a cold pooling catchment. Korean Journal of Agricultural and Forest Meteorology 14(4), 277-282. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2012.14.4.277
  3. Kim, S. O., and J. I. Yun, 2014: Improving usage of the Korea Meteorological Administration's digital forecasts in agriculture: III. Correction for advection effect on determination of daily maximum temperature over sloped surfaces. Korean Journal of Agricultural and Forest Meteorology 16(4), 297-303. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2014.16.4.297
  4. Kim, S. O., and J. I. Yun, 2015: Improving the usage of the Korea Meteorological Administration's digital forecasts in agriculture: IV. Estimation of daily sunshine duration and solar radiation based on 'Sky Condition' product. Korean Journal of Agricultural and Forest Meteorology 17(4), 281-289. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2015.17.4.281
  5. Kim, S. O., J. H. Kim, D. J. Kim, K. M. Shim, and J. I. Yun, 2015: Combined effects of wind and solar irradiance on the spatial variation of midday air temperature over a mountainous terrain. Asia-Pacific Journal of Atmospheric Sciences 51(3), 239-247. (in Korean with English abstract) https://doi.org/10.1007/s13143-015-0074-5
  6. Kim, S. O., and J. I. Yun, 2016: Feasibility of the lapse rate prediction at an hourly time interval. Korean Journal of Agricultural and Forest Meteorology 18(1), 55-63. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2016.18.1.55
  7. Kim, S. O., 2017a: Estimation of temporal surface air temperature under nocturnal inversion conditions. Korean Journal of Agricultural and Forest Meteorology 19(3), 75-85. (in Korean with English abstract) DOI: 10.5532/KJAFM.2017.19.3.75
  8. Kim, S. O., 2017b: Prediction of wind damage risk based on estimation of probability distribution of daily maximum wind speed. Korean Journal of Agricultural and Forest Meteorology 19(3), 130-139. (in Korean with English abstract) DOI: 10.5532/KJAFM.2017.19.3.130
  9. Kim D. J., S. O. Kim, J. H. Kim, and E. J. Yun, 2019: Establishment of geospatial schemes based on topo-climatology for farm-specific agrometeorological information. Korean Journal of Agricultural and Forest Meteorology 21(3), 117-128. (in Korean with English abstract) https://doi.org/10.5532/KJAFM.2019.21.3.117
  10. Konstantinos G. K., P. Busato, D. Moshou, S. Pearson, and D. Bochtis, 2018: Machine learning in agriculture: A review. Sensors 18, 2674. DOI: 10.3390/s18082674
  11. Rural Development Administration (RDA), 2018: Technical guide of early warning system for weather risk management in agricultural sector. 103pp. (in Korean)
  12. Mekonnen, Y., S. Namuduri, L. Burton, A. Sarwat, and S. Bhansali, 2020: Review - Machine learning techniques in wireless sensor network based precision agriculture. Journal of The Electrochemical Society 167, 037522. https://doi.org/10.1149/2.0222003jes
  13. WMO, 2000: Guidelines on performance assessment of public weather services. World Meteorological Organization Technical Document No. 1023, 62pp.
  14. Yun, J. I., S. O. Kim, J. H. Kim, and D. J. Kim, 2013: User-specific agrometeorological service to local farming community: A case study. Korean Journal of Agricultural and Forest Meteorology 15(4), 320-331. (in Korean with English abstract) DOI: 10.5532/KJAFM.2013.15.4.320
  15. Yun, J. I., 2014: Agrometeorological early warning system: A service infrastructure for climate-smart agriculture. Korean Journal of Agricultural and Forest Meteorology 16(4), 403-417. (in Korean with English abstract) DOI: 10.5532/KJAFM.2014.16.4.403
  16. Yun, J. I., 2015: A feasibility study of a field-specific weather service for small-scale farms in a topographically complex watershed. Korean Journal of Agricultural and Forest Meteorology 17(4), 317-325. DOI: 10.5532/KJAFM.2015.17.4.317
  17. Yun, E. J., and S. O. Kim, 2018: Estimation of hourly daytime air temperature on slope in complex terrain corrected by hourly solar radiation. Korean Journal of Agricultural and Forest Meteorology 20(4), 376-385. (in Korean with English abstract) DOI: 10.5532/KJAFM.2018.20.4.376