Abstract

Seasonal killing-frost frequency (KFF) during the cool/overwintering-crop growing season is important for the Canadian agricultural sector to prepare and respond to such extreme agrometeorological events. On the basis of observed daily surface air temperature across Canada for 1957–2007, this study found that more than 86% of the total killing-frost events occur in April–May and exhibit consistent variability over south-central Canada, the country’s major agricultural region. To quantify the KFF year-to-year variations, a simple index is defined as the mean KFF of the 187 temperature stations in south-central Canada. The KFF variability is basically dominated by two components: the decadal component with a peak periodicity around 11 yr and the interannual component of 2.5–3.8 yr. A statistical method called partial least squares (PLS) regression is utilized to uncover principal sea surface temperature (SST) modes in the winter preceding the KFF anomalies. It is found that most of the leading SST modes resemble patterns of El Niño–Southern Oscillation (ENSO) and/or the Pacific decadal oscillation (PDO). This indicates that ENSO and the PDO might be two dominant factors for the KFF variability. From a 41-yr training period (1957–97), a PLS seasonal prediction model is established, and 1-month-lead real-time forecasts are performed for the validation period of 1998–2007. A promising skill level is obtained. For the KFF variability, the prediction skill of the PLS model is comparable to or even better than the newly developed Canadian Seasonal to Interannual Prediction System (CanSIPS), which is a state-of-the-art global coupled dynamical system.

Abstract

This study investigates the North Atlantic Oscillation (NAO) on an intraseasonal time scale. The authors investigate the question of how the characteristics of NAO events are influenced by the choice of its definitions using daily NCEP–NCAR reanalysis data spanning 51 boreal winters. Four different NAO indexes are used in this study, including one station/gridpoint–based index and three pattern-based indexes.

It is found that the NAO events obtained using pattern–based indexes are quite similar to each other, while some notable differences are observed when the NAO is defined using the station/gridpoint–based index (NAO1). The characteristics of the pattern-based NAO are found to be more antisymmetric for its two phases, including its time-averaged spatial structures, its lifetime distributions, and time-evolving spatial structures. The NAO1, on the other hand, reveals some asymmetric characteristics between the two phases. Emphasis is placed on comparing the characteristics of the NAO events obtained using the NAO1 index and one of the pattern-based indices, that is, NAO2. The time-averaged spatial structures for the NAO2 expand across more of the polar region than the NAO1. The positive NAO1 shows a wave train signal over the Pacific–North American region during the setup phase, while the negative NAO1 is found to develop more locally over northern Europe and the North Atlantic. The wave activity flux for the NAO2 is primarily in the zonal direction while for the NAO1, on the other hand, it is mostly concentrated over the North Atlantic with a pronounced southward component. The barotropic vorticity equation is used to examine the physical mechanisms that drive the life cycle of the NAO.

Abstract

Based on the database of the Subseasonal to Seasonal (S2S) Prediction project of the World Weather Research Programme (WWRP)/World Climate Research Programme (WCRP), the influence of the North Atlantic Oscillation (NAO) on the Madden–Julian oscillation (MJO) and its forecast skill is investigated. It is found that most models can capture the MJO phase changes following positive and negative NAO events. About 20 days after initializing with a positive (negative) NAO, the forecast MJO appears most frequently in phase 7 (3), which corresponds to reduced (enhanced) convection in the tropical Indian Ocean and enhanced (suppressed) convection in the western Pacific. In most S2S models the MJO prediction skill is dependent on the NAO amplitude and phase in the initial condition. A strong NAO leads to a better MJO forecast skill than a weak NAO. The MJO skill tends to be higher when the forecast starts from a negative NAO than a positive NAO. These results indicates that there is a strong northern extratropical influence on the MJO and its forecast skill. It is important for numerical models to better represent the NAO influence to improve the simulation and prediction of the MJO.

Abstract

In this study, four machine-learning (ML) models [gradient boost decision tree (GBDT), light gradient boosting machine (LightGBM), categorical boosting (CatBoost), and extreme gradient boosting (XGBoost)] are used to perform seasonal forecasts for nonmonsoonal winter precipitation over the Eurasian continent (30°–60°N, 30°–105°E) (NWPE). The seasonal forecast results from a traditional linear regression (LR) model and two dynamic models are compared. The ML and LR models are trained using the data for the period of 1979–2010, and then these empirical models are used to perform the seasonal forecast of NWPE for 2011–18. Our results show that the four ML models have reasonable seasonal forecast skills for the NWPE and clearly outperform the LR model. The ML models and the dynamic models have skillful forecasts for the NWPE over different regions. The ensemble means of the forecasts including the ML models and dynamic models show higher forecast skill for the NWEP than the ensemble mean of the dynamic-only models. The forecast skill of the ML models mainly benefits from a skillful forecast of the third empirical orthogonal function (EOF) mode (EOF3) of the NWPE, which has a good and consistent prediction among the ML models. Our results also illustrate that the sea ice over the Arctic in the previous autumn is the most important predictor in the ML models in forecasting the NWPE. This study suggests that ML models may be useful tools to help improve seasonal forecasts of the NWPE.

Abstract

A simple GCM (SGCM) is constructed by adding empirically derived time-independent forcing terms to a dry primitive equation model. This yields a model with realistic time-mean jets and storm tracks. The SGCM is then used to study the equilibrium response to an imposed heating anomaly in the midlatitude Pacific, meant to represent an anomaly in the sea surface temperature. Using the SGCM’s own climatology as a basic state, the same model is then used to find the time-independent linear response to the same heating anomaly. The difference between the two responses is clearly attributed to the forcing due to anomalous transient eddies.

The sensitivity of the response to the strength and vertical profile of the heating, and to the presence of the wind speed in the surface flux parameterization, is explored. It is found that for a reasonable range of heating amplitude the transient eddy forcing is proportional to the heating and the responses to heating and cooling are almost antisymmetric. The antisymmetry breaks down at large amplitude. The vertical profile of heating has a small but systematic effect on the response: deeper heating leads to stronger equivalent barotropic features. The inclusion of wind speed in the surface flux parameterization alters the response mainly by virtue of altering the basic model climatology, rather than by any local effect on the heating.

The position of the heating anomaly is varied in both latitude and longitude to gain insight into the possible effects of systematic errors in GCMs. The time-independent linear response tends to move with the heating, but the eddy-driven nonlinear part remains relatively fixed and varies only in amplitude. The heating perturbation slightly modifies the first empirical orthogonal function of the model’s internal low frequency variability. The response projects strongly onto this pattern and the probability distribution function of the projection is significantly skewed.

Abstract

In this study, ensemble seasonal predictions of the Arctic Oscillation (AO) were conducted for 51 winters (1948–98) using a simple global atmospheric general circulation model. A means of estimating a priori the predictive skill of the AO ensemble predictions was developed based on the relative entropy (R) of information theory, which is a measure of the difference between the forecast and climatology probability density functions (PDFs). Several important issues related to the AO predictability, such as the dominant precursors of forecast skill and the degree of confidence that can be placed in an individual forecast, were addressed. It was found that R is a useful measure of the confidence that can be placed on dynamical predictions of the AO. When R is large, the prediction is likely to have a high confidence level whereas when R is small, the prediction skill is more variable. A small R is often accompanied by a relatively weak AO index. The value of R is dominated by the predicted ensemble mean. The relationship identified here, between model skills and the R of an ensemble prediction, offers a practical means of estimating the confidence level of a seasonal forecast of the AO using the dynamical model.

Through an analysis of the global sea surface temperature (SST) forcing, it was found that the winter AO-related R is correlated significantly with the amplitude of the SST anomalies over the tropical central Pacific and the North Pacific during the previous October. A large value of R is usually associated with strong SST anomalies in the two regions, whereas a poor prediction with a small R indicates that SST anomalies are likely weak in these two regions and the observed AO anomaly in the specific winter is likely caused by atmospheric internal dynamics.

Abstract

Multimodel ensemble (MME) seasonal forecasts are analyzed to evaluate numerical model performance in predicting the leading forced atmospheric circulation pattern over the extratropical Northern Hemisphere (NH). Results show that the time evolution of the leading tropical Pacific sea surface temperature (SST)-coupled atmospheric pattern (MCA1), which is obtained by applying a maximum covariance analysis (MCA) between 500-hPa geopotential height (Z ₅₀₀) in the extratropical NH and SST in the tropical Pacific Ocean, can be predicted with a significant skill in March–May (MAM), June–August (JJA), and December–February (DJF) one month ahead. However, most models perform poorly in capturing the time variation of MCA1 in September–November (SON) with 1 August initial condition. Two possible reasons for the models’ low skill in SON are identified. First, the models have the most pronounced errors in the mean state of SST and precipitation along the central equatorial Pacific. Because of the link between the divergent circulation forced by tropical heating and the midlatitude atmospheric circulation, errors in the mean state of tropical SST and precipitation may lead to a degradation of midlatitude forecast skill. Second, examination of the potential predictability of the atmosphere, estimated by the ratio of the total variance to the variance of the model forecasts due to internal dynamics, shows that the atmospheric potential predictability over the North Pacific–North American (NPNA) region is the lowest in SON compared to the other three seasons. The low ratio in SON is due to a low variance associated with external forcing and a high variance related to atmospheric internal processes over this area.

Abstract

The prediction skill of the North Atlantic Oscillation (NAO) in boreal winter is assessed in the operational models of the WCRP/WWRP Subseasonal-to-Seasonal (S2S) prediction project. Model performance in representing the contribution of different processes to the NAO forecast skill is evaluated. The S2S models with relatively higher stratospheric vertical resolutions (high-top models) are in general more skillful in predicting the NAO than those models with relatively lower stratospheric resolutions (low-top models). Comparison of skill is made between different groups of forecasts based on initial condition characteristics: phase and amplitude of the NAO, easterly and westerly phases of the quasi-biennial oscillation (QBO), warm and cold phases of ENSO, and phase and amplitude of the Madden–Julian oscillation (MJO). The forecasts with a strong NAO in the initial condition are more skillful than with a weak NAO. Those with negative NAO tend to have more skillful predictions than positive NAO. Comparisons of NAO skill between forecasts during easterly and westerly QBO and between warm and cold ENSO show no consistent difference for the S2S models. Forecasts with strong initial MJO tend to be more skillful in the NAO prediction than weak MJO. Among the eight phases of MJO in the initial condition, phases 3–4 and phase 7 have better NAO forecast skills compared with the other phases. The results of this study have implications for improving our understanding of sources of predictability of the NAO. The situation dependence of the NAO prediction skill is likely useful in identifying “windows of opportunity” for subseasonal to seasonal predictions.

Abstract

A simple GCM based on a primitive equation model with empirically derived time-independent forcing is used to make forecasts in the extended to seasonal range. The results are analyzed in terms of the response to a midlatitude Pacific sea surface temperature anomaly (SSTA), represented here by a heating perturbation. A set of 90-day, 30-member ensemble forecasts is made with 54 widely differing initial conditions, both with and without the SSTA. The development of the response, defined as the difference between ensemble means, is split into three 30-day averages: month 1, month 2, and month 3.

During month 1, ensemble members separate, and the local response and remote teleconnections are established. The local response is not very sensitive to the initial condition.

In month 2, the extended range, the responses are relatively strong and vary greatly from one initial condition to another. However, a linear analysis reveals that large variations in the response do not correlate strongly with large variations in the initial condition. The initial perturbations required to generate the observed variations in the response are relatively small, and may be difficult to isolate in a real forecasting situation.

In month 3, the seasonal range, variations between responses are much smaller. The initial condition loses its influence and the responses all start to resemble the equilibrium response discussed in Part I.

Abstract

The warm Arctic–cold continent pattern (WACC) of near-surface air temperature variability has often been associated with the connection between Arctic sea ice reduction and cold weather over the midlatitude continents. Whether the existence of this pattern is due to variability of sea ice or is caused by atmospheric internal dynamics is subject to debate. Based on a long integration of a primitive equation atmospheric model (SGCM), this study examines the origin of the warm Arctic–cold North American pattern (WACNA), which is characterized by a pair of opposite surface air temperature anomalies over the high-latitude Chukchi–Bering Sea region and the North American continent, in boreal winter on the intraseasonal time scale. The model atmosphere is maintained by a time-independent forcing, so that atmospheric internal dynamics is the only source of variability. It is found that the SGCM model simulates well the behavior of the observed WACNA pattern. The WACNA pattern develops by interacting with the time-mean flow and synoptic-scale transient eddies. Two pathways of Rossby wave propagation are associated with WACNA. The northern pathway originates from eastern Siberia moving eastward across the Bering Strait to Canada, and the southern pathway is rooted in the subtropical waveguide propagating across the eastern North Pacific. Our simulation of this pattern implies that tropospheric dynamics alone can generate the WACNA, and the predictability associated with this pattern is likely limited by its internal dynamics nature.