My research encompasses, but not limited to (i) Tropical Pacific variability; (ii) El Niño and Southern Oscillation (ENSO); (iii) Climatological mean state and its biases; (iv) ENSO teleconnections along coastal regions; (v) Ocean initialization for decadal climate predictions; (vi) Large scale ocean circulation as represented in multiple ocean state estimates (reanalyses) and reconciling their differences; and (vii) Influence of large-scale circulation on regional scale circulations along the coast. Few of my research papers are highlighted below.
Coupled model development in sub-seasonal to seasonal forecast system
Ray et al., 2023; Stefanova et al., 2022
Coupled model development towards building subseasonal-to-seasonal forecasting system involves developing land, ocean, waves, atmosphere, sea-ice components of the earth system model. Using advanced tools and statistical methods each prototype needs to be evaluated with observations to assess the improvements and degradations and suggest directions of improvement for successive prototypes generated iteratively. Due to its long memory ocean tends to retain its initial conditions on sub-seasonal (35 days) forecasts. However, coupled air-sea interactions influences near surface processes, particularly in regions of shallow thermocline depths. Thus SST forecasts at weeks 3-5 could be driven by changes in surface windstress that affects the thermocline depth, surface currents, and upper ocean mixing. On seasonal timescales (9 months), ocean plays a larger role in moving warm water volumes across regional ocean basins, thereby influencing the SST patterns. Ray et al., (2023) focuses on an experimental seasonal ensemble forecasting system in predicting the 2015/16 El Niño event. Factors contributing towards enhancing the predictions are presented through in-depth analysis of the physical process by which ocean-atmosphere coupling drives El Niño variability in the equatorial Pacific. Stefanova et al., (2022) provides a thorough assessment of the different components of the earth system in subseasonal forecast (35 day) prototypes.
Winter prior influence on following summer temperatures along N-CCS
Ray et al., 2022
Summer upwelling is a busy season for fisheries management along the Northern California Current System (N-CCS). Prior knowledge of ocean conditions from a season earlier facilitates fisheries management and marine ecosystem sustenance. Although it is well known that ENSO provide predictability to the following summer conditions at the surface, however is the same true for the subsurface? This study diagnoses seasonal to interannual variations in subsurface (below the mixed layer) processes linking summer temperatures at depth to winter prior. This connection is stronger during ENSO-neutral winters, indicating stronger predictability.
Re-emergence mechanism accounts for only 25% of the summer temperature variability at depth, leaving other sources of variations, such as oceanic advection to play a likely role. Investigating using an intermediate layer heat budget in layers below the mixed layer, the study exposes the crucial role of seasonally appearing poleward California Undercurrent in pre-conditioning the winter subsurface for the following summer. Arrival of coastally trapped waves and anomalous changes in CUC reduce the effects of this pre-conditioning during ENSO winters, which likely weakens the connection to the following summer temperatures.
Drivers of subsurface temperature in the Pacific Northwest
Oceanic teleconnections via coastally trapped waves (CTW) travel very fast along the US West coast providing added source of seasonal predictability. In this study the remotely generated CTW from equatorial Pacific during ENSO is shown to take around 2-4 months to influence the temperatures at depth in northern California Current System (N-CCS). Spicy waters via slow eastward currents in the North Pacific and poleward California Undercurrent (CUC) in the south also influence the temperatures at depth. CTW also has a strong impact on the depth of CUC, mostly following ENSO events. Seasonal variations in Pacific Decadal Oscillation (PDO) is shown to influence the subsurface temperature variations. The study quantifies the relative influence of each of these oceanic pathways with respect to individual time lags in driving the temperatures at depth in N-CCS. For the purpose a linear model with select predictors representing these remote oceanic processes explains almost 83% of the variability in temperatures at depth of the region.
The study highlights the large scale ocean circulations’ remote influence on Pacific-Northwest coast as diagnosed in a climate model that serves as boundary condition to a regionally downscaled ocean model (ROMS), used routinely for coastal seasonal predictions. Hence diagnosing and evaluating physical processes relevant to seasonal predictions in the climate model are relevant for enhanced reliable forecasts from the coastal model.
This work was highlighted by NOAA Pacific Marine Environmental Laboratory link
Drivers of equatorial Pacific cold tongue mean state
Using two layered approach in Ray et al., (2018a), the key role of Tropical Instability Waves (TIWs) induced vertical mixing is shown to maintain the mean state of equatorial cold tongue in the Pacific. The coupled model, GFDL-FLOR simulate weak TIWs, but has excessive stratifying effect from deep equatorial monthly-scale vertical advective cooling, which is induced by overly cyclonic off-equatorial wind stress curl. This effect is compensated by the weakened stratifying effects from TIWs and/or a strengthened destratifying effect of vertical mixing in the upper ocean. Compensating errors thus create a tortuous path toward better models, since ameliorating one dynamical bias may actually degrade overall model performance, until the formerly compensating bias(es) can be addressed as well.
Heat Budget for Equatorial cold tongue bias in coupled climate model
Current climate models not only vary widely in simulating ENSO variability, but also struggle to simulate the mean state, particularly the cold-dry bias in the eastern equatorial Pacific. To diagnose the source of this cold bias in a mixed layer (ML) heat budget using daily and monthly outputs from GFDL-FLOR and monthly output from reanalyses, the often used mixed layer (ML) of fixed depth ~50m is found to be too deep to emulate the heat budget from a monthly varying ML with a density criterion of 0.125 kg/m3. However a ML that is fixed temporally yet varying spatially provides a much better estimate of the ML heat budget. Using a two layered approach, monthly advective cooling is shown to balance less than half of the surface heating in ML, while the remaining heat that diffuses through the base of the ML below is balanced by the advective cooling in deeper layer.
ENSO variability from 1871-2007 – El Niño of two types!!
Giese et al., (2010); Giese and Ray (2011); Ray and Giese (2012)
El Niño and Southern Oscillation (ENSO) is associated with a weakening of the Walker circulation and appearance of anomalously warm water along the central-eastern equatorial Pacific. Given the recent strengthening of El Niño in warmer climate, it was predicted to intensify under greenhouse warming scenarios. How well do we understand ENSO variability? What was ENSO characteristics like, going back a century? These questions were answered using past records of SST (reconstructed statistically), as well as in an ocean reanalysis (using a dynamical ocean model) for the period of 1871-2007.
The studies show that the strength of El Niño and La Niña has strong decadal variability, with very strong El Niños occurring in both the early (1877/78, 1918/19) and late (1997/98) 20th century. Statistically, a 136 years record is too short to determine a change in El Niño frequency. A new metric (CHI – Center of Heat Index) was introduced and the intensity, location, duration and phase propagation of each events were determined individually. Unlike conventional ENSO indices (Niño 3.4, Niño 3, Niño 4, and Niño 1+2), CHI provided feasibility in tracking the anomaly across equatorial Pacific, giving an added dimension to the ENSO metric. The distribution of CHI locations during El Niño was indistinguishable from a Gaussian distribution, which clearly indicated the unlikelihood of EP-type and CP-type events being statistically distinguishable. Several follow-up studies further highlighted the diversity in El Niño location being a continuum across equatorial Pacific.