seasonmap field guide
The subseasonal desk: reading the six drivers together
Deterministic forecast skill decays with lead time and is largely gone beyond seven to ten days. Predictability at weeks two and three comes instead from slower components of the system — tropical convection, the stratospheric circulation, the angular-momentum budget — whose present state is observable and whose downstream effects arrive at documented lags. The Drivers panel tracks six such diagnostics. This page describes what each one measures, the thresholds that matter, and the order in which they are read.
Why these diagnostics extend predictability
The mid-latitude flow responds to persistent forcing. Anomalous tropical heating excites Rossby wave trains that arc into the extratropics over one to three weeks (Hoskins & Karoly 1981); stratospheric circulation anomalies descend and condition the tropospheric annular mode over two to six weeks (Baldwin & Dunkerton 2001). Because these forcings evolve more slowly than synoptic weather, their state today constrains the probability distribution of the pattern one to three weeks out — not any particular day's weather. Each instrument below monitors one such forcing.
1 · MJO phase wheel — RMM index (Wheeler & Hendon 2004)
The Madden–Julian Oscillation is the dominant mode of intraseasonal tropical variability: an envelope of enhanced convection propagating eastward at roughly 5 m s⁻¹ with a 30–60 day period. The wheel plots the RMM index; amplitude below 1 indicates no coherent event, and the phase number locates the enhanced convection. Composite CONUS impacts lag the phase by one to three weeks and are strongest in winter (see the MJO explainer for the phase table). The dashed extension is a VAR fit to the observed series — a statistical benchmark with useful skill to roughly 10–15 days, labeled as such.
2 · Equatorial moisture Hovmöller — convectively coupled waves (Wheeler & Kiladis 1999)
Precipitable-water departures averaged 10°S–10°N, by longitude and time. Two propagation speeds are distinguishable by eye: the MJO envelope near 5 m s⁻¹ (a broad band advancing ~4° of longitude per day) and convectively coupled Kelvin waves at 12–25 m s⁻¹ (narrow streaks crossing 10–20° per day). Kelvin-wave passages modulate convection and measurably raise the probability of tropical cyclogenesis in their convectively active phase for several days (Schreck 2015). The operational question this diagram answers: whether tropical convection over the next one to two weeks will be reinforced or suppressed at a given longitude.
3 · Rossby wave-train Hovmöller — v250, 35–60°N
Meridional wind at 250 hPa, band-averaged and plotted by longitude and time. Wave packets appear as alternating northerly/southerly couplets sloping eastward; because group velocity exceeds phase velocity, energy propagates downstream faster than individual troughs move — the mechanism of downstream development (Chang 1993). Packets routinely cross the Pacific in five to seven days. This is the transit diagnostic: forcing identified on instruments 1, 2, 4, or 6 should subsequently appear here as a propagating packet, with its arrival longitude and date readable from the slope.
4 · East Asian mountain torque — surface pressure × orography, 25–50°N 70–140°E
The zonal pressure force exerted on the Tibetan Plateau and surrounding terrain, integrated to a torque about Earth's axis. Positive East Asian torque events — cold surface high pressure building against the plateau — transfer angular momentum to the atmosphere and are followed, at lags near one week, by East Asian cold surges and extension of the Pacific jet (Iskenderian & Salstein 1998; Lott, Robertson & Ghil 2004), with downstream CONUS pattern changes roughly a week after that. The index here is computed directly from GFS analyses; departures from the running mean are the operative signal.
5 · 10 hPa zonal wind at 60°N — the SSW criterion (Nov–Mar)
The strength of the stratospheric polar vortex. The standard definition of a major sudden stratospheric warming is the reversal of this wind from westerly to easterly during winter (Charlton & Polvani 2007); major SSWs occur in roughly six of ten Northern Hemisphere winters. Following a reversal, easterly anomalies descend over days to weeks, and the surface response — negative Arctic Oscillation, high-latitude blocking, elevated cold-outbreak risk — is concentrated in the following two to six weeks (Baldwin & Dunkerton 2001), though roughly a third of events couple weakly to the surface. The dashed GFS trajectory extends the gauge 16 days, so a model-forecast reversal is visible at that lead. Outside November–March the vortex does not exist and the gauge reads climatological summer easterlies; the panel notes this explicitly. Background: the polar vortex explainer.
6 · Cyclone phase space — Hart (2003), per active storm
Each NHC active storm, tracked through the GFS forecast and plotted in the two parameters that classify cyclone structure: B, the storm-motion-relative 900–600 hPa thickness asymmetry (B > 10 m indicates a frontal, asymmetric system), and −VT, the thermal-wind measure of core structure (positive: warm core). A trajectory from the warm-core/symmetric quadrant toward cold-core/frontal is extratropical transition. ET matters beyond the storm itself: recurving, transitioning cyclones inject heat and momentum into the jet and are a documented source of downstream ridge–trough amplification and reduced hemispheric predictability (Jones et al. 2003) — following an ET event, instrument 3 frequently shows the resulting wave packet.
Reading order
The instruments are read tropics-outward, because that is the direction of causality on these timescales:
- Tropical state (1, 2): MJO amplitude, phase, and projection; Kelvin-wave activity. This sets the base forcing for the next one to three weeks.
- Momentum and storm triggers (4, 6): recent torque spikes; any storm forecast to transition. Both are impulsive events that excite wave trains.
- Transit (3): whether wave packets consistent with those sources are actually propagating, and where they arrive.
- Stratospheric state (5, winter): vortex strength and forecast trend, which conditions the amplitude and persistence of whatever the troposphere is doing.
- Cross-check: compare the implied evolution against the long-range model pattern and the medium-range ensembles, and weight confidence by the number of independent diagnostics in agreement. Where they conflict, the appropriate forecast is closer to climatology.
A documented sequence
The chain most relevant to winter CONUS forecasting, with its approximate lags: an MJO event reaching the West Pacific (phases 6–7) is followed within about a week by East Asian cold-air outbreaks and a positive mountain-torque event; the associated Pacific jet extension launches a Rossby wave packet that crosses to North America in five to seven days, favoring downstream amplification in the 10–15 day window. If the attendant upward wave flux disturbs the stratospheric vortex, the 10 hPa gauge weakens or reverses, and the surface influence of that disruption persists for the following two to six weeks. Each step is observable on this panel at the time it occurs, at the instrument number given above.
Baldwin, M. P., & Dunkerton, T. J. (2001). Stratospheric harbingers of anomalous weather regimes. Science, 294, 581–584.
Chang, E. K. M. (1993). Downstream development of baroclinic waves as inferred from regression analysis. J. Atmos. Sci., 50, 2038–2053.
Charlton, A. J., & Polvani, L. M. (2007). A new look at stratospheric sudden warmings. J. Climate, 20, 449–469.
Hart, R. E. (2003). A cyclone phase space derived from thermal wind and thermal asymmetry. Mon. Wea. Rev., 131, 585–616.
Hoskins, B. J., & Karoly, D. J. (1981). The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci., 38, 1179–1196.
Iskenderian, H., & Salstein, D. A. (1998). Regional sources of mountain torque variability and high-frequency fluctuations in atmospheric angular momentum. Mon. Wea. Rev., 126, 1681–1694.
Jones, S. C., et al. (2003). The extratropical transition of tropical cyclones. Wea. Forecasting, 18, 1052–1092.
Lott, F., Robertson, A. W., & Ghil, M. (2004). Mountain torques and Northern Hemisphere low-frequency variability. J. Atmos. Sci., 61, 1259–1271.
Schreck, C. J. (2015). Kelvin waves and tropical cyclogenesis. Mon. Wea. Rev., 143, 3996–4011.
Wheeler, M., & Kiladis, G. N. (1999). Convectively coupled equatorial waves. J. Atmos. Sci., 56, 374–399.
Wheeler, M. C., & Hendon, H. H. (2004). An all-season real-time multivariate MJO index. Mon. Wea. Rev., 132, 1917–1932.