DOI QR코드

DOI QR Code

Evaluating the prediction models of leaf wetness duration for citrus orchards in Jeju, South Korea

제주 감귤 과수원에서의 이슬지속시간 예측 모델 평가

  • Park, Jun Sang (Applied Meteorology Research Division, National Institute of Meteorological Sciences) ;
  • Seo, Yun Am (Applied Meteorology Research Division, National Institute of Meteorological Sciences) ;
  • Kim, Kyu Rang (Applied Meteorology Research Division, National Institute of Meteorological Sciences) ;
  • Ha, Jong-Chul (Applied Meteorology Research Division, National Institute of Meteorological Sciences)
  • 박준상 (국립기상과학원 응용기상연구과) ;
  • 서윤암 (국립기상과학원 응용기상연구과) ;
  • 김규랑 (국립기상과학원 응용기상연구과) ;
  • 하종철 (국립기상과학원 응용기상연구과)
  • Received : 2018.08.23
  • Accepted : 2018.09.13
  • Published : 2018.09.30

Abstract

Models to predict Leaf Wetness Duration (LWD) were evaluated using the observed meteorological and dew data at the 11 citrus orchards in Jeju, South Korea from 2016 to 2017. The sensitivity and the prediction accuracy were evaluated with four models (i.e., Number of Hours of Relative Humidity (NHRH), Classification And Regression Tree/Stepwise Linear Discriminant (CART/SLD), Penman-Monteith (PM), Deep-learning Neural Network (DNN)). The sensitivity of models was evaluated with rainfall and seasonal changes. When the data in rainy days were excluded from the whole data set, the LWD models had smaller average error (Root Mean Square Error (RMSE) about 1.5hours). The seasonal error of the DNN model had the similar magnitude (RMSE about 3 hours) among all seasons excluding winter. The other models had the greatest error in summer (RMSE about 9.6 hours) and the lowest error in winter (RMSE about 3.3 hours). These models were also evaluated by the statistical error analysis method and the regression analysis method of mean squared deviation. The DNN model had the best performance by statistical error whereas the CART/SLD model had the worst prediction accuracy. The Mean Square Deviation (MSD) is a method of analyzing the linearity of a model with three components: squared bias (SB), nonunity slope (NU), and lack of correlation (LC). Better model performance was determined by lower SB and LC and higher NU. The results of MSD analysis indicated that the DNN model would provide the best performance and followed by the PM, the NHRH and the CART/SLD in order. This result suggested that the machine learning model would be useful to improve the accuracy of agricultural information using meteorological data.

2016년부터 2017년까지 제주 감귤과수원 11개 지점에서 관측한 기상 및 이슬 자료를 이용하여 이슬지속시간 예측 모델을 평가하였다. 이슬지속시간 모델의 민감도와 예측 정확도 평가에는 4 가지 모델(Number of Hours of Relative Humidity, Classification And Regression Tree/Stepwise Linear Discriminant, Penman-Monteith, Deep-learning Neural Network)이 사용되었다. 모델의 민감도는 강우와 계절 변화에 따라 평가하였다. 전체 자료에서 강우일 자료를 제외하면 이슬지속시간 모델들은 평균 오차(평균제곱근오차 약 1.5 hours)가 적게 나타났다. 기계학습 모델은 겨울을 제외한 계절별 오차가 비슷한 크기(평균제곱근오차 약 3 hours)로 나타났다. 나머지 모델들은 여름에 오차(평균제곱근오차 약 9.6 hours)가 가장 크고 겨울에 가장 작은 것(평균제곱근오차 약 3.3 hours)으로 나타났다. 모델 예측 정확도 평가 방법은 통계적 오차 분석 방법과 평균 제곱 편차 회귀 분석 방법을 사용하였다. 통계오차를 통한 모델 성능은 DNN 모델이 가장 우수한 반면에 CART/SLD 모델은 예측 정확도가 가장 낮게 나타났다. 평균제곱 편차(MSD)는 모델의 선형성을 세 가지(제곱 바이어스(SB), 비균일성 기울기(NU), 상관관계 부족(LC)) 구성요소로 구분하여 분석하는 방법이다. 모델 성능이 우수할수록 SB와 LC는 감소하였고 NU는 증가하는 경향이 나타났다. MSD 분석 결과 DNN 모델이 가장 우수하였으며 다음으로 PM, NHRH, CART/SLD 순으로 나타났다. 본 연구에서 활용된 기계학습 모델은 기상 정보를 이용한 다른 농업정보 생산의 정확도 개선에 크게 기여할 것으로 판단된다.

Keywords

References

  1. Allen, R. G., L. S. Pereira, D. Raes, and M. Smith, 1998: Crop evapotranspiration: Guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper No. 56, Rome, Italy: United Nations FAO.
  2. Beysens, D., 2016: Estimationg dew yield worldwide from a few meteo data. Atmospheric Research 167, 146-155. https://doi.org/10.1016/j.atmosres.2015.07.018
  3. Beysens, D., M. Muselli, V. Nikolayev, R. Narhe, and I. Milimouk, 2005: Measurement and modeling of dew in island coastal and alpine areas. Atmospheric Research 73, 1-22. https://doi.org/10.1016/j.atmosres.2004.05.003
  4. Cosh, M. H., E. D. Kabela, B. Hornbuckle, M. L. Gleason, T. J. Jackson, and J. H. Prueger, 2009: Observation of dew amount using in situ and satellite measurements in an agricultural landscape. Agricultural and Forest Meteorology 149, 1082-1086. https://doi.org/10.1016/j.agrformet.2009.01.004
  5. Gauch, H. G., J. T. Hwang, and G. W. Fick, 2003: Model evaluation by comparison of model-based predictions and measured values. Agronomy Journal 95, 1442-1446. https://doi.org/10.2134/agronj2003.1442
  6. Gleason, M. L., S. E. Taylor, T. M. Loughin, and K. J. Koehler, 1994: Development and validation of an empirical model to estimate the duration of dew periods. Plant disease 78, 1011-1016.
  7. Gleason, M. L., 2007: Validation of weather inputs for disease warning systems. American Phytopathological Society 97(7), 147.
  8. Hinton, G. E., S. Osindero, and Y. W. Teh, 2006: A fast learning algorithm for deep belief nets. Neural computation 18(7), 1527-1554. https://doi.org/10.1162/neco.2006.18.7.1527
  9. Hong, K. O., M. S. Suh, and D. K. Rha, 2006: Temporal and spatial variations of precipitation in South Korea for recent 30 years (1976-2005) and geographic environments. Journal of Korean Earth Sciences Society 27(4), 433-449.
  10. Kim, K. S., S. E. Taylor, M. L. Gleason, R. Villalobos, and L. F. Arauz, 2005: Estimation of leaf wetness duration using empirical models in northwestern Costa Rica. Agricultural and Forest Meteorology 129, 53-67. https://doi.org/10.1016/j.agrformet.2004.11.009
  11. Kim, K. S., S. E. Taylor, M. L. Gleason, and K. J. Koehler, 2002: Model to enhance site-specific estimation of leaf wetness duration. Plant disease 86, 179-185. https://doi.org/10.1094/PDIS.2002.86.2.179
  12. Kim, S. S., S. M. Jang, H. J. Baek, H. Y. Choi, and W. T. Kwon, 2006: Climatological variability of temperature and precipitation in Jeju. Journal of Korean Earth Sciences Society 27(2), 188-197.
  13. Klemm, O., C. Milford, M. A. Sutton, G. Spindler, and E. Van Putten, 2002: A climatology of leaf surface wetness. Theoretical and Applied Climatology 71, 107-117. https://doi.org/10.1007/s704-002-8211-5
  14. LeCun, Y., Y. Bengio, and G. Hinton, 2015: Deep learning. nature 521(7553), 436-444. https://doi.org/10.1038/nature14539
  15. Lee, H., J. B. Jee, J. S. Min, S. Kim, and J. H. Chae, 2018: Analysis of meteorological and radiation characteristics using WISE observation data. Journal of Korean Earth Sciences Society 39(1), 89-102. https://doi.org/10.5467/JKESS.2018.39.1.89
  16. Lee, Y. J., and J. Nang, 2016: A personal video event classification method based on multi-modalities by DNN-Learning. Journal of KIISE 43(11), 1281-1297. https://doi.org/10.5626/JOK.2016.43.11.1281
  17. Luo, W., and J. Goudriaan, 2004: Estimating dew formation in rice, using seasonally averaged dial patterns of weather variables. NJAS wageningen journal of life sciences 51, 391-406. https://doi.org/10.1016/S1573-5214(04)80004-6
  18. Maestre-Valero, J. F., R. Ragab, V. Martinez-Alvarez, and A. Baille, 2012: Estimation of dew yield from radiative condensers by means of an energy balance model. Journal of Hydrology 460, 103-109.
  19. Park, J. H., D. H. Shin, and C. B. Kim, 2017: Deep learning model for electric power demand prediction using special day separation and prediction elements extension. Journal of Advanced Navigation Technology 21(4), 365-370. https://doi.org/10.12673/JANT.2017.21.4.365
  20. Park, J. S., K. R. Kim, M. Kang, and B. J. Kim, 2017: The influence of evaporation from a stream on fog events in the middle Nakdong River. Journal of Korean Earth Sciences Society 38(6), 395-404. https://doi.org/10.5467/JKESS.2017.38.6.395
  21. Rowlandson, T., M. Gleason, P. Sentelhas, T. Gillespie, C. Thomas, and B. Hornbuckle, 2015: Reconsidering leaf wetness duration determination for plant disease management. Plant Disease 99, 310-319. https://doi.org/10.1094/PDIS-05-14-0529-FE
  22. Sentelhas, P. C., T. J. Gillespie, M. L. Gleason, J. E. B. Monteiro, J. R. M. Pezzopane, and M. J. Pedro, 2006: Evaluation of a Penman-Monteith approach to provide "reference" and crop canopy leaf wetness duration estimates. Agricultural and Forest Meteorology 141, 105-117. https://doi.org/10.1016/j.agrformet.2006.09.010
  23. Sentelhas, P. C., A. Dalla Marta, S. Orlandini, E. A. Santos, T. J. Gillespie, and M. L. Gleason, 2008: Suitability of relative humidity as an estimator of leaf wetness duration. Agricultural and forest meteorology 148, 392-400. https://doi.org/10.1016/j.agrformet.2007.09.011
  24. Shin, D. H., and C. B. Kim, 2018: Short term forecast model for solar power generation using RNN-LSTM. Journal of Advanced Navigation Technology 22(3), 233-239. (in Korean) https://doi.org/10.12673/JANT.2018.22.3.233
  25. Montone, V. O., C. W. Fraisse, N. A. Peres, P. C. Sentelhas, M. Gleason, M. Ellis, and G. Schnabel, 2016: Evaluation of leaf wetness duration models for operational use in strawberry disease-warning systems in four US states. International journal of biometeorology 60, 1761-1774. https://doi.org/10.1007/s00484-016-1165-4
  26. Veronica, N. S., W. F. Clyde, A. P. Natalia, C. C. James, and C. Amy, 2010: Spatial variability of leaf wetness duration in citrus canopies. Proceedings of the Florida State Horticultural Society 123, 49- 55.
  27. Wallach, D., and B. Goffinet, 1989: Mean squared error of prediction as a criterion for evaluating and comparing system models. Ecological Modeling 44, 299-306. https://doi.org/10.1016/0304-3800(89)90035-5
  28. Wieland, R., W. Mirschel, K. Groth, A. Pechenick, and K. Fukuda, 2011: A new method for semi-automatic fuzzy training and its application in environmental modeling. Environmental modeling & software 26, 1568-1573. https://doi.org/10.1016/j.envsoft.2011.07.017