This programme of research is intended to make climate simulations more consistent both with the multi-scale nature of climate, and with related scaling symmetries of the partial differential equations which govern climate. This will be achieved by moving away from the traditional deterministic approach to the closure problem in computational fluid dynamics, and towards a more novel description of physical processes near and below the truncation scale of climate models, using contemporary nonlinear stochastic-dynamic mathematics. The aim of the proposed research is to produce the world's first Probabilistic Earth System Model. The consequences are enormous: a comprehensive climate model with reduced biases against observations, a model which will be capable of producing estimates of uncertainty in its own predictions, and a model which can make use of emerging energy-efficient probabilistic processor hardware, key to practical success as we approach the era of the exascale supercomputer. The development of the prototype Probabilistic Earth-System Model will open a new era of international scientific collaboration on climate model development, and has the potential to influence climate policy, on mitigation, adaptation and on geoengineering, at the highest governmental and intergovernmental levels.
Towards the Prototype Probabilistic Earth-System Model for Climate Prediction [2014 - 2017] (Funded by ERC)
EASM-3: Quantifying Predictability Limits, Uncertainties, Mechanisms, and Regional Impacts of Pacific Decadal Climate Variability [2014 - 2017] (co-PI, Funded by NSF)
The project team proposes a coordinated research effort to better understand the basic physical dynamics of Pacific decadal variability and assess the skill of Pacific decadal predictability, along with its uncertainties and practical value. The research focuses on Community Earth System Model (CESM), with its vast repository of archived runs supplemented with targeted predictability experiments. The analysis focuses on using sophisticated statistical models (Linear Inverse Models) to identify statistical relations among variables, diagnose physical processes, and isolate potentially predictable components of the flows. It also involves using regional coupled atmosphere-ocean, along with uncoupled ocean and atmosphere models, to enhance the understanding of regional response and its potential for practical use in forecasting. The project brings together scientists skilled with developing decadal climate diagnostics, making both statistical and dynamical predictions, and executing regional coupled climate downscaling and regional high-resolution ocean modeling.
A Nudging and Ensemble Forecasting Approach to Identify and Correct Tropical Pacific Bias-Producing Processes in CESM [2014 - 2017] (PI, Funded by NOAA)
We plan to study the spatiotemporal structures of bias development in CESM forecasts, launched from numerous initial states and during which random ENSO and MJO events occur, to determine the relative importance of poor mean-state representation versus the integrated impacts of the transient flows. This bias development will be studied as a function of season to account for significant changes in the background state of the coupled oceanatmosphere system in the tropical Pacific. We will also seek to ascribe these effects to wellknown physical processes for the specific climate modes of variability. We will test the sensitivity of the bias development to changes in coupled model resolution and model parameter selection. We will also implement nudging experiments (towards observations) to pinpoint where the worst parts of the biases develop apart from the nudged variables.
Assessing the Impact of Diurnal Wind Variability [2014 - 2018] (co-I, Funded by NASA)
The planned launch of RapidScat aboard the International Space Station, which does not follow a sun synchronous orbit, will offer further opportunities to assess diurnal variability of ocean winds. The objective of this study will focus on first characterizing seasonal and interannual variability of diurnal winds both using multi-sensor measurements (from ASCAT, OSCAT, and WindSat) and also RapidScat measurements once they become available. This will allow us to characterize seasonal changes in the diurnal amplitude and year-to-year variations in the structure and magnitude of diurnal winds. Weather stations and mooring-based winds will be used to validate satellite-derived results. While satellite-based assessments of diurnal winds have typically been limited to examining a single diurnally varying harmonic, the more extensive sampling should make it possible to consider semi-diurnal and higher frequencies as well. The study will also use an one-dimensional upper-ocean model with high vertical resolution to quantify the role that diurnal winds play in upper-ocean processes. Results will be used to develop improvements for a more idealized model used to represent upper ocean processes in prognostic climate models. Our goals are to assess the impact of including or neglecting diurnally varying winds on large-scale ocean circulation studies and to quantify the impact of the rectified effects of radiative forcing and diurnal winds on surface temperature and salinity, upper ocean heat content and air-sea heat fluxes.
Evaluating the Roles of Factors Critical to MJO Simulations Using the NCAR CAM3 with Deterministic and Stochastic Convection Parameterization Closures (Funded by NSF [2012 - 2014])
The Madden-Julian Oscillation (MJO) is a large-scale pattern of precipitation and atmospheric circulation anomalies which forms over the Indian Ocean and propagates slowly eastward towards the central equatorial Pacific Ocean. Impacts of the MJO are felt over much of the earth, and the presence of the MJO in the central equatorial Pacific Ocean exerts a strong modulation of hurricane activity in the Gulf of Mexico. Work conducted for this project is intended to improve the representation of the MJO in global climate models (GCMs), particularly the National Center for Atmospheric Research Community Atmosphere Model (CAM). This work was being done in collaboration with Dr. Guang Zhang and Dr. Mitch Moncrieff. The questions addressed in the research focus on specific factors believed to be essential to a successful MJO simulation:
Simulation and data assimilation of South East Pacific (SEP) ocean mesoscale structure (Funded by NSF [2008 - 2012])The ability of high resolution eddy-resolving ocean models to accurately resolve the South East Pacific eddy structure was critically tested using the VOCALS-REx (VAMOS Ocean-Cloud-Atmosphere-Land Study Regional Experiment) oceanographic mesoscale survey and satellite data. We are using ROMS (Regional Ocean Modeling System) to simulate the horizontal and vertical eddy structure. Synthetic analyses for the Regional Experiment using ocean data assimilation to understand the eddy heat/freshwater/tracer/nutrient transports from the coastal upwelling region to the remote SEP, and interactions between the eddies, the mixed layer and the deeper ocean. I am currently testing data assimilation twin experiments in Incremental Strong-constraint 4DVAR (IS4DVAR) framework of a high resolution (7 km horizontal resolution) regional ocean model of the South East Pacific region under the guidance of Prof. Art Miller and Prof. Bruce Cornuelle. I have run sample experiments to test the corrections in the ocean state estimation due to data assimilation in twin experiments. I plan to use ship cruise data from the VOCALS-REx funded cruise in Oct-08 to assimilate real time data from satellites and ship cruise surveys to better estimate the ocean state for the period during the ship cruise.
Publications from this project:
Inverse Methods for Data assimilation in nonlinear problems (Funded by ONR [2008-2010])We are also learning and extending the application of optimal nonlinear filtering theory for ocean state estimation in collaboration with Prof. Ibrahim Hoteit, Prof. Bruce Cornuelle and Prof. Art Miller. We are testing different flavors of Ensemble Kalman Filters and also particle filters to use in simple nonlinear problems such as Lorenz-86 model and then extend the study to simple ocean models such as the barotropic vorticity model. We plan to work on the development, implementation and testing of several Particle Kalman Filters in simple test cases to analyze their behavior and understand their merits and demerits in data assimilation for real ocean and atmospheric models.
Publications from this project:
Subramanian et al. 2011
Coupled Ocean-Atmospheric processes in MJO-ENSO Dynamics (Funded by ONR [2009 - 2013])Coupled oceanic-atmospheric processes were analyzed using a regional coupled model (SCOAR) in the Indian Ocean region to understand the controlling parameters on the Intraseasonal Oscillations of the Asian Summer Monsoon (ASM). This work is in collaboration with Prof. Art Miller, Prof. Raghu Murtugudde, Dr. Markus Jochum and Dr. Hyodae Seo. We are analyzing the influences of the air-sea fluxes and its variability on the intraseasonal variability of the ASM. Also, research interests diversify into the variability in the intraseasonal timescale in the internal ocean like the mixed layer and its predictability on the variability of the ASM at the intraseasonal timescale.
Publications from this project: