The impact of the quasi‐biennial oscillation (QBO) on the prediction of tropical intraseasonal convection, including the Madden Julian Oscillation (MJO) and Boreal Summer Intraseasonal Oscillation (BSISO), is assessed in the WMO Subseasonal to Seasonal (S2S) forecast database using the real‐time OLR based MJO (ROMI) index. It is shown that the ROMI prediction skill for the boreal winter MJO, measured by the maximum time at which the anomaly correlation coefficient exceeds 0.6, is higher by 5 to 10 days in the QBO easterly phase than its westerly phase. This difference occurs even in models with low tops and poorly resolved stratospheres. MJO predictability, as measured by signal to noise ratio in the S2S ensemble, also shows a similar difference between the two QBO phases, and results from a simple linear regression model show consistent behavior as well. Analysis of the ROMI index derived from observations indicates that the MJO is more coherent and stronger in the QBO easterly phase than in the westerly phase. These results suggest that the skill dependence on QBO phase results from the initial state of the MJO, the regularity of its propagation in the verifying observations, or most likely a combination of the two, but not on an actual stratospheric influence on the MJO within the model simulations. In contrast to the robust QBO‐MJO connection in boreal winter, the BSISO prediction skill exhibited by the S2S models in boreal summer is greater in the QBO
Few studies have utilized machine learning techniques to predict or understand the Madden‐Julian oscillation (MJO), a key source of subseasonal variability and predictability. Here, we present a simple framework for real‐time MJO prediction using shallow artificial neural networks (ANNs). We construct two ANN architectures, one deterministic and one probabilistic, that predict a real‐time MJO index using maps of tropical variables. These ANNs make skillful MJO predictions out to ∼18 days in October‐March and ∼11 days in April‐September, outperforming conventional linear models and efficiently capturing aspects of MJO predictability found in more complex, dynamical models. The flexibility and explainability of simple ANN frameworks are highlighted through varying model input and applying ANN explainability techniques that reveal sources and regions important for ANN prediction skill. The accessibility, performance, and efficiency of this simple machine learning framework is more broadly applicable to predict and understand other Earth system phenomena.
more » « less- NSF-PAR ID:
- 10368357
- Publisher / Repository:
- DOI PREFIX: 10.1029
- Date Published:
- Journal Name:
- Journal of Advances in Modeling Earth Systems
- Volume:
- 14
- Issue:
- 5
- ISSN:
- 1942-2466
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract westerly phase than in theeasterly phase during the 1999 to 2010 period. This is consistent with the observation that BSISO OLR anomalies are stronger in the QBO westerly phase during that period. However, this relationship between the QBO and BSISO in boreal summer changes in recent decades: BSISO is weaker in QBO westerly than easterly during 1979–2000. Correspondingly, the QBO impact on BSISO prediction in boreal summer also reverses in that period as well in a statistical model, whereas this statistical model shows a consistent QBO impact on MJO prediction in boreal winter over the past four decades. -
more » « less
Abstract The Madden‐Julian oscillation (MJO) is the leading source of global subseasonal predictability; however, many dynamical forecasting systems struggle to predict MJO propagation through the Maritime Continent. Better understanding the biases in simulated physical processes associated with MJO propagation is the key to improve MJO prediction. In this study, MJO prediction skill, propagation processes, and mean state biases are evaluated in reforecasts from models participating in the Subseasonal Experiment (SubX) and Subseasonal to Seasonal (S2S) prediction projects. SubX and S2S reforecasts show MJO prediction skill out to 4.5 weeks based on the Real‐time Multivariate MJO index consistent with previous studies. However, a closer examination of these models' representation of MJO propagation through the Maritime Continent reveals that they fail to predict the MJO convection, associated circulations, and moisture advection processes beyond 10 days with most of models underestimating MJO amplitude. The biases in the MJO propagation can be partly associated with the following mean biases across the Indo‐Pacific: a drier low troposphere, excess surface precipitation, more frequent occurrence of light precipitation rates, and a transition to stronger precipitation rates at lower humidity than in observations. This indicates that deep convection occurs too frequently in models and is not sufficiently inhibited when tropospheric moisture is low, which is likely due to the representation of entrainment.
-
more » « less
Abstract The composite structure of the Madden–Julian oscillation (MJO) has long been known to feature pronounced Rossby gyres in the subtropical upper troposphere, whose existence can be interpreted as the forced response to convective heating anomalies in the presence of a subtropical westerly jet. The question of interest here is whether these forced gyre circulations have any subsequent effects on divergence patterns in the tropics and the Kelvin-mode component of the MJO. A nonlinear spherical shallow water model is used to investigate how the introduction of different background jet profiles affects the model’s steady-state response to an imposed MJO-like stationary thermal forcing. Results show that a stronger jet leads to a stronger Kelvin-mode response in the tropics up to a critical jet speed, along with stronger divergence anomalies in the vicinity of the forcing. To understand this behavior, additional calculations are performed in which a localized vorticity forcing is imposed in the extratropics, without any thermal forcing in the tropics. The response is once again seen to include pronounced equatorial Kelvin waves, provided the jet is of sufficient amplitude. A detailed analysis of the vorticity budget reveals that the zonal-mean zonal wind shear plays a key role in amplifying the Kelvin-mode divergent winds near the equator, with the effects of nonlinearities being of negligible importance. These results help to explain why the MJO tends to be strongest during boreal winter when the Indo-Pacific jet is typically at its strongest.
Significance Statement The MJO is a planetary-scale convectively coupled equatorial disturbance that serves as a primary source of atmospheric predictability on intraseasonal time scales (30–90 days). Due to its dominance and spontaneous recurrence, the MJO has a significant global impact, influencing hurricanes in the tropics, storm tracks, and atmosphere blocking events in the midlatitudes, and even weather systems near the poles. Despite steady improvements in subseasonal-to-seasonal (S2S) forecast models, the MJO prediction skill has still not reached its maximum potential. The root of this challenge is partly due to our lack of understanding of how the MJO interacts with the background mean flow. In this work, we use a simple one-layer atmospheric model with idealized heating and vorticity sources to understand the impact of the subtropical jet on the MJO amplitude and its horizontal structure.
-
more » « less
Abstract Predicting Pacific Decadal Oscillation (PDO) transitions and understanding the associated mechanisms has proven a critical but challenging task in climate science. As a form of decadal variability, the PDO is associated with both large‐scale climate shifts and regional climate predictability. We show that artificial neural networks (ANNs) predict PDO persistence and transitions with lead times of 12 months onward. Using layer‐wise relevance propagation to investigate the ANN predictions, we demonstrate that the ANNs utilize oceanic patterns that have been previously linked to predictable PDO behavior. For PDO transitions, ANNs recognize a build‐up of ocean heat content in the off‐equatorial western Pacific 12–27 months before a transition occurs. The results support the continued use of ANNs in climate studies where explainability tools can assist in mechanistic understanding of the climate system.
-
Explaining machine learning models with interactive natural language conversations using TalkToModelmore » « less
Abstract Practitioners increasingly use machine learning (ML) models, yet models have become more complex and harder to understand. To understand complex models, researchers have proposed techniques to explain model predictions. However, practitioners struggle to use explainability methods because they do not know which explanation to choose and how to interpret the explanation. Here we address the challenge of using explainability methods by proposing TalkToModel: an interactive dialogue system that explains ML models through natural language conversations. TalkToModel consists of three components: an adaptive dialogue engine that interprets natural language and generates meaningful responses; an execution component that constructs the explanations used in the conversation; and a conversational interface. In real-world evaluations, 73% of healthcare workers agreed they would use TalkToModel over existing systems for understanding a disease prediction model, and 85% of ML professionals agreed TalkToModel was easier to use, demonstrating that TalkToModel is highly effective for model explainability.
