Atmospheric vapor pressure deficit (VPD) is a useful proxy for meteorological fire-prone conditions in the southwestern United States (SWUS). Here, we demonstrate that the leading VPD mode, with a pronounced decadal component, displays a center-of-action in the SWUS. In its dry phase, this mode is characterized by reduced soil moisture, high-pressure systems, and La Niña-like conditions, and connected to decadal sea surface temperature (SST) variability in the Pacific and Atlantic basins. Leveraging climate simulations, a backward model-analog technique reveals persistent SST signals a few seasons preceding the dry VPD mode. While state-of-the-art models can reasonably capture the spatial patterns of VPD, challenges emerge in representing the decadal component of the leading VPD mode, which predominantly arises from misrepresentations of the actual rather than saturation vapor pressure.

VPD measures the difference between saturation and actual vapor pressure, and its variability is closely related to fire activity in the western US. Here, we assess the forecast skill of monthly VPD variability using a state-of-the-art dynamical forecast system and statistical predictions, such as the persistence forecast and model-analog forecasts. In the model-analog framework, we select analog states resembling the observed initial conditions from the model space, and the subsequent evolution of those initial model-analogs yields forecast ensembles. 

Dynamical forecasts demonstrate skillful predictions of VPD variability in the western US, exceeding the persistence forecast skill, which indicates additional sources of VPD predictability within the climate system. To quantify the contribution of different climate variables to VPD prediction, we develop a weighted model-analog forecast and evaluate its skill in comparison to VPD-only and unweighted forecasts. Our findings suggest that sea surface temperature is a critical source of VPD predictability over the western US. The optimally weighted model-analog exhibits forecast skill for VPD variability comparable to that of the dynamical forecast system.

Diagnosing El Niño-Southern Oscillation (ENSO) predictability within operational forecast models is hindered by computational expense and the need for initialization with three-dimensional fields generated by global data assimilation. We instead examine multi-year ENSO predictability since the late 1800s using the model-analog technique, which has neither limitation. We first draw global coupled model states from pre-industrial control simulations, from the Coupled Model Intercomparison Project Phase 6, that are chosen to initially match observed monthly sea surface temperature and height anomalies in the Tropics. Their subsequent 36-month model evolution are the hindcasts, whose 20th century ENSO skill is comparable to twice-yearly hindcasts generated by a state-of-the-art European operational forecasting system. Despite the so-called spring predictability barrier, present throughout the record, there is substantial second-year ENSO skill, especially after 1960. Overall, ENSO exhibited notably high values of both amplitude and skill towards the end of the 19th century, and again in recent decades.

While Pacific climate variability is largely understood based on El Niño-Southern Oscillation (ENSO), the North Pacific focused Pacific decadal oscillation and the basin-wide interdecadal Pacific oscillation, the role of the South Pacific, including atmospheric drivers and cross-scale interactions, has received less attention. Using reanalysis data and model outputs, here we propose a paradigm for South Pacific climate variability whereby the atmospheric Pacific-South American (PSA) mode acts to excite multiscale spatiotemporal responses in the upper South Pacific Ocean. We find the second mid-troposphere PSA pattern is fundamental to stochastically generate a mid-latitude sea surface temperature quadrupole pattern that represents the optimal precursor for the predictability and evolution of both the South Pacific decadal oscillation and ENSO several seasons in advance. We find that the PSA mode is the key driver of oceanic variability in the South Pacific subtropics that generates a potentially predictable climate signal linked to the tropics.

By employing a first-order autoregressive model, we demonstrate that the atmospheric Pacific-South American patterns 1 and 2 (PSA1 and PSA2), defined as the second and third empirical orthogonal function (EOF) modes of the monthly 500hPa geopotential height anomalies in the Southern Hemisphere, act as stochastic drivers for the first two sea surface temperature (SST) modes. These modes are identified as the South Pacific Decadal Oscillation (SPDO) and the SST quadrupole pattern.

The South Pacific decadal oscillation (SPDO) characterizes the Southern Hemisphere contribution to the Pacific-wide interdecadal Pacific oscillation (IPO) and is analogous to the Pacific decadal oscillation (PDO) centered in the North Pacific. In this study, upper ocean variability and potential predictability of the SPDO is examined in HadISST data and an atmosphere-forced ocean general circulation model. The potential predictability of the IPO-related variability is investigated in terms of both the fractional contribution made by the decadal component in the South, tropical and North Pacific Oceans and in terms of a doubly integrated first-order autoregressive (AR1) model. Despite explaining a smaller fraction of the total variance, we find larger potential predictability of the SPDO relative to the PDO. We identify distinct local drivers in the western subtropical South Pacific, where nonlinear baroclinic Rossby wave–topographic interactions act to low-pass filter decadal variability. In particular, we show that the Kermadec Ridge in the southwest Pacific enhances the decadal signature more prominently than anywhere else in the Pacific basin. Applying the doubly integrated AR1 model, we demonstrate that variability associated with the Pacific–South American pattern is a critically important atmospheric driver of the SPDO via a reddening process analogous to the relationship between the Aleutian low and PDO in the North Pacific—albeit that the relationship in the South Pacific appears to be even stronger. Our results point to the largely unrecognized importance of South Pacific processes as a key source of decadal variability and predictability.

A multivariate linear inverse model (LIM) is developed to demonstrate the mechanisms and seasonal predictability of the dominant modes of variability from the tropical and South Pacific Oceans. We construct a LIM whose covariance matrix is a combination of principal components derived from tropical and extratropical sea surface temperature, and South Pacific Ocean vertically averaged temperature anomalies. Eigen-decomposition of the linear deterministic system yields stationary and/or propagating eigenmodes, of which the least damped modes resemble El Niño–Southern Oscillation (ENSO) and the South Pacific decadal oscillation (SPDO). We show that although the oscillatory periods of ENSO and SPDO are distinct, they have very close damping time scales, indicating that the predictive skill of the surface ENSO and SPDO is comparable. The most damped noise modes occur in the midlatitude South Pacific Ocean, reflecting atmospheric eastward-propagating Rossby wave train variability. We argue that these ocean wave trains occur due to the high-frequency atmospheric variability of the Pacific–South American pattern imprinting onto the surface ocean. The ENSO spring predictability barrier is apparent in LIM predictions initialized in March–May (MAM) but displays a significant correlation skill of up to ~3 months. For the SPDO, the predictability barrier tends to appear in June–September (JAS), indicating remote but delayed influences from the tropics. We demonstrate that subsurface processes in the South Pacific Ocean are the main source of decadal variability and further that by characterizing the upper ocean temperature contribution in the LIM, the seasonal predictability of both ENSO and the SPDO variability is increased.

A stochastically forced linear inverse model (LIM) of the combined modes of variability from the tropical and South Pacific Oceans is used to investigate the linear growth of optimal initial perturbations and to identify the spatiotemporal features of the stochastic forcing associated with the atmospheric Pacific–South American patterns 1 and 2 (PSA1 and PSA2). Optimal initial perturbations are shown to project onto El Niño–Southern Oscillation (ENSO) and South Pacific decadal oscillation (SPDO), where the inclusion of subsurface South Pacific Ocean temperature variability significantly increases the multiyear linear predictability of the deterministic system. We show that the optimal extratropical sea surface temperature (SST) precursor is associated with the South Pacific meridional mode, which takes from 7 to 9 months to linearly evolve into the final ENSO and SPDO peaks in both the observations and as simulated in an atmosphere-forced ocean model. The optimal subsurface precursor resembles its peak phase, but with a weak amplitude, representing oceanic Rossby waves in the extratropical South Pacific. The stochastic forcing is estimated as the residual by removing the deterministic dynamics from the actual tendency under a centered difference approximation. The resulting stochastic forcing time series satisfies the Gaussian white noise assumption of the LIM. We show that the PSA-like variability is strongly associated with stochastic SST forcing in the tropical and South Pacific Oceans and contributes not only to excite the optimal initial perturbations associated with ENSO and the SPDO but in general to activate the entire stochastic SST forcing, especially in austral summer.