Performance Comparison of High Resolution Weather Research and Forecasting Model Output with North American Mesoscale Model Initialization Grid Forecasts

Raby, John; Passner, Jeff; Vaucher, Gail; Raby, Yasmina

doi:10.21236/ada561957

Cited by 1 publication

(2 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model results were comparable or better than efforts from previous ML frost forecast studies (Verdes et al, 2000;Diedrichs et al, 2018) (24-h F1 = 0.39-0.68), although a direct comparison between studies is difficult due to differences in methodology, input data, environmental settings of where the models were applied, and prediction lead-time. Our temperature predictions also achieved better performance metrics than available large-scale numerical weather models such as the HRRR model and compared to models assessed in previous studies (Raby et al, 2012). Such forecasts could provide accurate predictions of frost at the point scale in areas of complex topography where numerical models perform relatively poorly.…”

Section: Discussionmentioning

confidence: 54%

“…Due to the localized nature of temperature in Northern New Mexico and the course spatial resolution of alternative temperature predictions (1 km-WRF, 3 km-HRRR, 12 km-NAMeso), the ML model predictions are likely an improvement over existing data available to farmers. The WRF and NAMeso models were found by one study to have a regional average RMSE of 2.6–2.88°C for a 24-h lead time (Raby et al, 2012 ), which is slightly worse than the model results for the Alcalde station (RMSE= 2.19–2.37°C). Further, The ML models presented outperformed the HRRR model forecast by a large margin, resulting in an improved 6-h forecast RMSE of about 1.5°C.…”

Section: Discussionmentioning

confidence: 93%

See 1 more Smart Citation

Frost prediction using machine learning and deep neural network models

Talsma

Solander²,

Mudunuru³

et al. 2023

Front. Artif. Intell.

View full text Add to dashboard Cite

This study describes accurate, computationally efficient models that can be implemented for practical use in predicting frost events for point-scale agricultural applications. Frost damage in agriculture is a costly burden to farmers and global food security alike. Timely prediction of frost events is important to reduce the cost of agricultural frost damage and traditional numerical weather forecasts are often inaccurate at the field-scale in complex terrain. In this paper, we developed machine learning (ML) algorithms for the prediction of such frost events near Alcalde, NM at the point-scale. ML algorithms investigated include deep neural network, convolution neural networks, and random forest models at lead-times of 6–48 h. Our results show promising accuracy (6-h prediction RMSE = 1.53–1.72°C) for use in frost and minimum temperature prediction applications. Seasonal differences in model predictions resulted in a slight negative bias during Spring and Summer months and a positive bias in Fall and Winter months. Additionally, we tested the model transferability by continuing training and testing using data from sensors at a nearby farm. We calculated the feature importance of the random forest models and were able to determine which parameters provided the models with the most useful information for predictions. We determined that soil temperature is a key parameter in longer term predictions (>24 h), while other temperature related parameters provide the majority of information for shorter term predictions. The model error compared favorable to previous ML based frost studies and outperformed the physically based High Resolution Rapid Refresh forecasting system making our ML-models attractive for deployment toward real-time monitoring of frost events and damage at commercial farming operations.

show abstract

Section: Discussionmentioning

confidence: 54%

Section: Discussionmentioning

confidence: 93%