Search Results

You are looking at 1 - 8 of 8 items for

  • Author or Editor: Yuval x
  • Refine by Access: All Content x
Clear All Modify Search
Yuval

Abstract

Neural network (NN) training is the optimization process by which the relation between the NN input and output is established. A new formulation for the NN training is presented where an NN model is reconstructed such that it produces predicted output data optimally fitting the observed ones. The optimal level of fit is determined by minimization of the generalized cross-validation function, which is integrated in the training. The training process is fully automated, does not require the user to set aside data for validation, and enables objective testing and evaluation of the predictions. Results are demonstrated and discussed using synthetic data produced by Lorenz’s low-order circulation model and on real data from the equatorial Pacific.

Full access
Yuval

Abstract

A procedure to enhance neural network (NN) predictions of tropical Pacific sea surface temperature anomalies and calculating their estimated errors is presented. A simple linear correction enables more accurate predictions of warm and cold events but can result in introduction of larger errors in other cases. The prediction error estimates aid recognizing erroneously magnified anomalies and are used to sort the predictions into El Niño, La Niña, and neutral states. The error estimation process is based on bootstrap resamplings of the data and construction of a large number of bootstrap prediction replicas. A statistic calculated on the set of bootstrap replicas that corresponds to each of the actual predictions is used to estimate the prediction’s errors. The method is demonstrated on NN prediction of the Niño-3.4 index.

Full access
Yuval Yevnin
and
Yaron Toledo

Abstract

The paper presents a combined numerical–deep learning (DL) approach for improving wind and wave forecasting. First, a DL model is trained to improve wind velocity forecasts by using past reanalysis data. The improved wind forecasts are used as forcing in a numerical wave forecasting model. This novel approach, used to combine physics-based and data-driven models, was tested over the Mediterranean. The correction to the wind forecast resulted in ∼10% RMSE improvement in both wind velocity and wave height over reanalysis data. This significant improvement is even more substantial at the Aegean Sea when Etesian winds are dominant, improving wave height forecasts by over 35%. The additional computational costs of the DL model are negligible compared to the costs of either the atmospheric or wave numerical model by itself. This work has the potential to greatly improve the wind and wave forecasting models used nowadays by tailoring models to localized seasonal conditions, at negligible additional computational costs.

Significance Statement

Wind and wave forecasting models solve a set of complicated physical equations. Improving forecasting accuracy is usually achieved by using a higher-resolution, empirical coefficients calibration or better physical formulations. However, measurements are rarely used directly to achieve better forecasts, as their assimilation can prove difficult. The presented work bridges this gap by using a data-driven deep learning model to improve wind forecasting accuracy, and the resulting wave forecasting. Testing over the Mediterranean Sea resulted in ∼10% RMSE improvement. Inspecting the Aegean Sea when the Etesian wind is dominant shows an outstanding 35% improvement. This approach has the potential to improve the operational atmospheric and wave forecasting models used nowadays by tailoring models to localized seasonal conditions, at negligible computational costs.

Free access
Yuval
and
William W. Hsieh

Abstract

A novel neural network (NN)–based scheme performs nonlinear model output statistics (MOS) for generating precipitation forecasts from numerical weather prediction (NWP) model output. Data records from the past few weeks are sufficient for establishing an initial MOS connection, which then adapts itself to the ongoing changes and modifications in the NWP model. The technical feasibility of the algorithm is demonstrated in three numerical experiments using the NCEP reanalysis data in the Alaskan panhandle and the coastal region of British Columbia. Its performance is compared with that of a conventional NN-based nonadaptive scheme. When the new adaptive method is employed, the degradation in the precipitation forecast skills due to changes in the NWP model is small and is much less than the degradation in the performance of the conventional nonadaptive scheme.

Full access
Janni Yuval
and
Yohai Kaspi

Abstract

The relation between the mean meridional temperature gradient and eddy fluxes has been addressed by several eddy flux closure theories. However, these theories give little information on the dependence of eddy fluxes on the vertical structure of the temperature gradient. The response of eddies to changes in the vertical structure of the temperature gradient is especially interesting since global circulation models suggest that as a result of greenhouse warming, the lower-tropospheric temperature gradient will decrease whereas the upper-tropospheric temperature gradient will increase. The effects of the vertical structure of baroclinicity on atmospheric circulation, particularly on the eddy activity, are investigated. An idealized global circulation model with a modified Newtonian relaxation scheme is used. The scheme allows the authors to obtain a heating profile that produces a predetermined mean temperature profile and to study the response of eddy activity to changes in the vertical structure of baroclinicity. The results indicate that eddy activity is more sensitive to temperature gradient changes in the upper troposphere. It is suggested that the larger eddy sensitivity to the upper-tropospheric temperature gradient is a consequence of large baroclinicity concentrated in upper levels. This result is consistent with a 1D Eady-like model with nonuniform shear showing more sensitivity to shear changes in regions of larger baroclinicity. In some cases, an increased temperature gradient at lower-tropospheric levels might decrease the eddy kinetic energy, and it is demonstrated that this might be related to the midwinter minimum in eddy kinetic energy observed above the northern Pacific.

Full access
Janni Yuval
and
Yohai Kaspi

Abstract

The atmosphere exhibits two distinct types of jets: the thermally driven subtropical jet and the more poleward eddy-driven jet. Depending on location and season, these jets are often merged or separated, and their position, structure, and intensity strongly influence the eddy fields. Here, the authors study the sensitivity of eddies to changes in the jets’ amplitudes and positions in an idealized general circulation model. A modified Newtonian relaxation scheme that has a very short relaxation time for the mean state and a long relaxation time for eddies is used. This scheme makes it possible to obtain any zonally symmetric temperature distribution and is used to systematically modify the jets’ amplitudes and locations. It is found that eddies are more sensitive to changes in the amplitude of the eddy-driven jet than to changes in the amplitude of the subtropical jet. Furthermore, when the eddy-driven jet is shifted poleward, eddies tend to intensify. These results are tested for robustness in two different reference simulations: one resembling a situation where the subtropical and eddy-driven jets are nearly merged and one when they are separated.

Full access
Janni Yuval
and
Yohai Kaspi

Abstract

Motivated by the expectation that under global warming upper-level meridional temperature gradients will increase while lower-level temperature gradients will decrease, the relations between the vertical structure of baroclinicity and eddy fields are investigated. The sensitivity of eddies and the relation between the mean available potential energy and eddy quantities are studied for cases where the vertical structure of the lapse rate and meridional temperature gradient are modified. To investigate this systematically, an idealized general circulation model with a Newtonian cooling scheme that has a very short relaxation time for the mean state and a long relaxation time for eddies is used. This scheme allows for any chosen zonally mean state to be obtained with good precision. The results indicate that for similar change in the lapse rate or meridional temperature gradient, eddies are more sensitive to changes in baroclinicity where it is already large. Furthermore, when the vertical structure of the lapse rate or the meridional temperature gradient is modified, there is no universal linear relation between the mean available potential energy and eddy quantities.

Full access
Janni Yuval
and
Yohai Kaspi

Abstract

Global warming projections show an anomalous temperature increase both at the Arctic surface and at lower latitudes in the upper troposphere. The Arctic amplification decreases the meridional temperature gradient, and simultaneously decreases static stability. These changes in the meridional temperature gradient and in the static stability have opposing effects on baroclinicity. The temperature increase at the upper tropospheric lower latitudes tends to increase the meridional temperature gradient and simultaneously increase static stability, which have opposing effects on baroclinicity as well. In this study, a dry idealized general circulation model with a modified Newtonian cooling scheme, which allows any chosen zonally symmetric temperature distribution to be simulated, is used to study the effect of Arctic amplification and lower-latitude upper-level warming on eddy activity. Due to the interplay between the static stability and meridional temperature gradient on atmospheric baroclinicity changes, and their opposing effect on atmospheric baroclinicity, it is found that both the Arctic amplification and lower-latitude upper-level warming could potentially lead to both decreases and increases in eddy activity, depending on the exact prescribed temperature modifications. Therefore, to understand the effect of global warming–like temperature trends on eddy activity, the zonally symmetric global warming temperature projections from state-of-the-art models are simulated. It is found that the eddy kinetic energy changes are dominated by the lower-latitude upper-level warming, which tends to weaken the eddy kinetic energy due to increased static stability. On the other hand, the eddy heat flux changes are dominated by the Arctic amplification, which tends to weaken the eddy heat flux at the lower levels.

Open access