We review prediction efforts of El Niño events in the tropical Pacific with particular focus on using modern machine learning (ML) methods based on artificial neural networks. With current classical prediction methods using both statistical and dynamical models, the skill decreases substantially for lead times larger than about 6 months. Initial ML results have shown enhanced skill for lead times larger than 12 months. The search for optimal attributes in these methods is described, in particular those derived from complex network approaches, and a critical outlook on further developments is given.
We apply Gaussian density neural network and quantile regression neural network ensembles to predict the El Niño–Southern Oscillation. Both models are able to assess the predictive uncertainty of the forecast by predicting a Gaussian distribution and the quantiles of the forecasts, respectively. This direct estimation of the predictive uncertainty for each given forecast is a novel feature in the prediction of the El Niño–Southern Oscillation by statistical models. The predicted mean and median, respectively, show a high‐correlation skill for long lead times (r=0.5, 12 months) for the 1963–2017 evaluation period. For the 1982–2017 evaluation period, the probabilistic forecasts by the Gaussian density neural network can better estimate the predictive uncertainty than a standard method to assess the predictive uncertainty of statistical models.
Abstract. Atmosphere models with resolutions of several tens of kilometres take subgrid-scale variability in the total specific humidity q t into account by using a uniform probability density function (PDF) to predict fractional cloud cover. However, usually only mean relative humidity, RH, or mean clear-sky relative humidity, RH cls , is used to compute hygroscopic growth of soluble aerosol particles. While previous studies based on limited-area models and also a global model suggest that subgrid-scale variability in RH should be taken into account for estimating radiative forcing due to aerosol-radiation interactions (RFari), here we present the first estimate of RFari using a global atmospheric model with a parameterization for subgrid-scale variability in RH that is consistent with the assumptions in the model. For this, we sample the subsaturated part of the uniform RH-PDF from the cloud cover scheme for its application in the hygroscopic growth parameterization in the ECHAM6-HAM2 atmosphere model. Due to the non-linear dependence of the hygroscopic growth on RH, this causes an increase in aerosol hygroscopic growth. Aerosol optical depth (AOD) increases by a global mean of 0.009 (∼ 7.8 % in comparison to the control run). Especially over the tropics AOD is enhanced with a mean of about 0.013. Due to the increase in AOD, net top of the atmosphere clear-sky solar radiation, SW net,cls , decreases by −0.22 W m −2 (∼ −0.08 %). Finally, the RFari changes from −0.15 to −0.19 W m −2 by about 31 %. The reason for this very disproportionate effect is that anthropogenic aerosols are disproportionally hygroscopic.
Abstract. Atmosphere models with resolutions of several tens of kilometres take subgrid-scale variability of the total specific humidity q t into account by using a uniform probability density function (PDF) to predict fractional cloud cover. However, usually only mean relative humidity RH or mean clear-sky relative humidity RH cls is used to compute hygroscopic growth of soluble aerosol particles. In this study, a stochastic parameterization of subgrid-scale variability of RH cls is applied. For this, we sample the subsaturated part of the uniform RH-PDF from the cloud cover scheme for application in association with the
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.