Method of Rain Attenuation Prediction Based on Long–Short Term Memory Network

Cornejo, Andrés; Landeros‐Ayala, Salvador; Matías, J. M.; Ortíz, Flor; Martínez, Raquel; Natera, Miguel Alejandro Salas

doi:10.1007/s11063-022-10749-1

Cited by 10 publications

(5 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are three gate units imported into the LSTM: the historical information can be selectively retained by the forget gate; the input gate is employed to pick new input information; and the output of the hidden layer is determined by the output gate. Based on these features, LSTM has been used to classify dry and wet periods (Habi & Messer, 2018), estimate attenuation (Cornejo et al., 2022; Pu, Liu, & He, 2020), and nowcast rainfall fields (Xian et al., 2020). In this paper, LSTM cells are used to learn the time dependence of spatial rainfall distribution characteristics.…”

Section: Methodsmentioning

confidence: 99%

Reconstructing and Nowcasting the Rainfall Field by a CML Network

Zhang,

Liu,

Zou

2023

Earth and Space Science

View full text Add to dashboard Cite

Currently, the opportunistic method to estimate rainfall using commercial microwave links (CMLs) has been shown as an efficient way to complement traditional instruments in terms of spatial‐temporal resolution and coverage. In this paper, we collected data from 26 CMLs in Jiangyin City, Jiangsu Province, and conducted experiments on rainfall field reconstruction and nowcasting. First, the raw CML data were processed to invert the path‐averaged rainfall intensity. Second, the algorithms of inverse distance weighting (IDW) and ordinary kriging (OK) interpolation were employed to reconstruct the rainfall field. Then a 10‐min prediction of the rainfall field was achieved using a nowcasting model based on the long short‐term memory neural network and a setup window was introduced to improve the prediction performance of the first few minutes. The reconstruction results show that the average correlation coefficient (ACC) and the average root mean square error (ARMSE) between the IDW‐based results and daily cumulative rainfall from rain gauges (RGs) are 0.89 and 8.69 mm, respectively, while the ACC and ARMSE between the OK‐based results and RG data are 0.89 and 9.13 mm, respectively. The nowcasting results show that the ACC between the prediction results with a 5‐min setup window and the IDW‐retrieved rainfall fields can reach 0.91 at the first minute and gradually decrease to 0.20 within 10 min. Furthermore, the model has better nowcasting performance for stratiform precipitation and mixed precipitation compared to convective precipitation.

show abstract

Section: Methodsmentioning

confidence: 99%

Reconstructing and Nowcasting the Rainfall Field by a CML Network

Zhang,

Liu,

Zou

2023

Earth and Space Science

View full text Add to dashboard Cite

show abstract

“…For each combination of the hyperparameters, models were trained for at most 800 epochs (unless prevented by the early stopping procedure). We adopted two metrics to evaluate the model on the test data: the R 2 score, a relative metric derived from the MSE, and the mean absolute error (MAE), as it is common practice in regression problems [5,37,51].…”

Section: Neural Predictive Modelmentioning

confidence: 99%

“…More recent research has successfully applied neural networks and data mining approaches to predict hurricanes' trajectories [34,35]. Furthermore, deep neural networks and random forests were shown to be the most promising machine learning model used in meteorological nowcasting [36,37].…”

Section: Introductionmentioning

confidence: 99%

Forecasting Convective Storms Trajectory and Intensity by Neural Networks

Borghi,

Guariso,

Sangiorgio

2024

Forecasting

View full text Add to dashboard Cite

Convective storms represent a dangerous atmospheric phenomenon, particularly for the heavy and concentrated precipitation they can trigger. Given their high velocity and variability, their prediction is challenging, though it is crucial to issue reliable alarms. The paper presents a neural network approach to forecast the convective cell trajectory and intensity, using, as an example, a region in northern Italy that is frequently hit by convective storms in spring and summer. The predictor input is constituted by radar-derived information about the center of gravity of the cell, its reflectivity (a proxy for the intensity of the precipitation), and the area affected by the storm. The essential characteristic of the proposed approach is that the neural network directly forecasts the evolution of the convective cell position and of the other features for the following hour at a 5-min temporal resolution without a relevant loss of accuracy in comparison to predictors trained for each specific variable at a particular time step. Besides its accuracy (R2 of the position is about 0.80 one hour in advance), this machine learning approach has clear advantages over the classical numerical weather predictors since it runs at orders of magnitude more rapidly, thus allowing for the implementation of a real-time early-warning system.

show abstract

“…In [30], a back-propagation neural network has been used to predict the rain rate and corresponding attenuation, however it has not been tested for higher frequency bands where channel fluctuations are significantly higher. In [31], synthetic rain data is generated and utilized to train a long short-term memory (LSTM) neural network. The model used provided promising results in terms of rain attenuation prediction.…”

Section: A Related Workmentioning

confidence: 99%

Deep Learning Forecasting and Statistical Modeling for Q/V-Band LEO Satellite Channels

Homssi¹,

Chan²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>This paper presents a practical approach for Q/V-band modeling for low Earth orbit satellite channels based on tools from machine learning and statistical modeling. The developed Q/V-band LEO satellite channel model is presented in two folds; (i) a real-time forecasting method using model-based deep learning, intended for real-time operation of satellite terminals, and (ii) a statistical channel simulator that generates the path loss as a time-series random process, intended for system design and research. The provided approach capitalizes on real satellite measurements that are obtained from AlphaSat's Q/V-band transmitter at different geographic latitudes, to model the radio channel. The results show that model-based deep learning forecasting can outperform conventional statistically derived prediction methods for varying rain and elevation angle profiles. Moreover, it can also provide more accuracy in long-term prediction in comparison to current state-of-the-art machine learning approaches for radio channel prediction. Results for the statistical channel simulator is shown to produce synthetic radio excess path loss values for varying satellite passes by capitalizing on empirical statistical models obtained from real measurements.</p>

show abstract

Method of Rain Attenuation Prediction Based on Long–Short Term Memory Network

Cited by 10 publications

References 43 publications

Reconstructing and Nowcasting the Rainfall Field by a CML Network

Reconstructing and Nowcasting the Rainfall Field by a CML Network

Forecasting Convective Storms Trajectory and Intensity by Neural Networks

Deep Learning Forecasting and Statistical Modeling for Q/V-Band LEO Satellite Channels

Contact Info

Product

Resources

About