Funded Projects

Program (s): Climate Variability and Predictability

Competition:

Award Number: | View Publications on Google Scholar

Despite the key role of precipitation in climate and climate impacts, it remains one of the most poorly modeled climate variables. In addition to well-known biases in the simulated climatology of tropical precipitation, there are also biases in tropical precipitation sensitivity to climate perturbations. For example, even if a model has its convection zone in the proper mean location vis a vis the observations, it does not necessarily follow that the sensitivity of the convection to variations in temperature, wind, or inflow water vapor is correct. Under the previous grant, we have developed a number of tools that we propose can contribute to identifying and addressing the biases in convective processes. In particular, we outline new diagnostic methods to understand the transition to strong convection, presenting preliminary examples using satellite observations of precipitation and column water vapor. We propose to apply these diagnostics to better constrain the temperature and moisture dependences of the onset of precipitation required for climate model convective parameterizations, and to contrast these measures in observations to current model simulations. We propose to complement the satellite observations with in situ sounding data for vertical structure in strongly convecting regions, beginning with the high temporal coverage Atmospheric Radiation Measurement (ARM) Program site at Nauru Island. We will focus first on quantifying the convective threshold and its dependences on thermodynamic variables. The water vapor-temperature dependence of the onset of convection can be important to mechanisms by which large-scale processes, such as the wave dynamics occurring in teleconnections, or inflow air masses from dry regions into the convective margins, interact with the small-scale convective processes. We propose to examine how variations in inflow from the dry Southeastern Pacific into the South Pacific Convergence Zone (SPCZ) interact with the threshold for the the onset of convection in models compared to observations. This will help quantify the contribution of this interaction to the mean and sensitivity biases of the convectivemargin in this region. We propose to analyze similar impacts of the variations in the moisture relative to the convective threshold in ENSO variations. We will aim to formulate our diagnostics of observed and modeled convective processes in terms that can be directly useful to modeling groups. 

Program (s): Climate Variability and Predictability

Competition: Improved Understanding of Tropical Pacific Processes, Biases, and Climatology

Award Number: NA14OAR4310274 | View Publications on Google Scholar

Moist dynamical processes, originating in the atmosphere but involving ocean-atmosphere feedbacks, are among the leading effects requiring better constraints to address tropical Pacific biases and many applications to Pacific variability. We propose to bring together two themes developed under prior work: fast-process diagnostics from observations and parameter perturbation runs aimed at assessing sensitivity of moist processes to parameterized physics. Specifically, we will have available from prior work a set of runs with the Community Earth System Model (CESM1; atmospheric component Community Atmosphere Model 5) that perturb the convective physics in both uncoupled and coupled modes. These include nonstandard, high time-resolution output to aid assessment of fast-process diagnostics and sufficient length to establish statistical significance in quantities that might elude short-term forecast experiments. Data from a related set of perturbed physics runs from the NOAA Geophysical Fluid Dynamics Laboratory High Resolution Atmospheric Model (HIRAM) model will also be available via existing collaborations. In both models, initial results suggest high parameter sensitivity in the tropical Pacific, for instance, precipitation differences exceeding ± 3 mm/day across large parts of the basin for convection-related parameters varied across their feasible range. The proposed work addresses some of the challenges in making use of such information: (i) Sensitivity does not necessarily equate to improvement. We will quantify contributions to this across multiple variables and parameters, including assessing trade-offs where some metrics improve while others degrade for a given parameter change. (ii) The impacts on the climatology involve large-scale dynamical ocean-atmosphere feedbacks even in experiments where the parameterization change is known. We will aim to disentangle such effects using hypothesis-driven investigation informed by simpler models. Examples of this include assessment of convective instability as a function of parameter and model state using a column version of the CAM, and convective margins diagnostics. (iii) Reduction of biases should not simply be a tuning exercise based on improvement in the climatology. Rather, we will seek cases where current fast-process diagnostics can provide independent constraints on the parameter range or parameterization form. For example, diagnostics for convective onset as a function of temperature and water vapor will be used to constrain the entrainment range. Processes exhibiting high sensitivity in the Pacific will be used to target the development of further diagnostics, and parallels with common error modes in the Coupled Model Intercomparison Project phase 5 will be examined.

The proposed work, for the competition Improved Understanding of Tropical Pacific Processes, Biases, and Climatology (Earth System Science Competition 3, Climate Variability and Predictability), addresses the following aspects of the competition goals: intercomparison of model parameterizations, including convection and clouds, and reduced and conceptual modeling coordinated with analysis of full coupled model experiments and development of metrics for atmospheric relationships that constrain moist dynamical processes key to understanding and reducing biases. It addresses elements of the NOAA long-term climate goal as described in NOAA’s Next-Generation Strategic Plan in the core capability of “Understanding and modeling”, supporting the “ Predictions and projections” capability. Furthermore, the information from the observational analysis will help to design and refine the diagnostics that can be obtained from observing systems. The proposed work is relevant to societal challenges identified in the NGSP of climate impacts in water resources, changes in extremes of weather and climate, and provides current climate baselines for model projections for mitigating climate change impacts.

Program (s): Climate Variability and Predictability

Competition: Understanding Arctic Sea Ice Mechanisms and Predictability

Award Number: NA15OAR4310165 | View Publications on Google Scholar

The shrinking extent and thickness of the Arctic sea ice cover, as well as the major loss of multi-year pack ice is allowing greater access to the Arctic. In order to make use of this new accessibility efficiently and to guarantee safe operations, high-resolution sea ice forecasts are required on a variety of time scales, from hours to days, months, and seasons. Currently, shortcomings in our modeling capability preclude accurate prediction of ice characteristics on the necessary variety of temporal and spatial scales. The ability to simulate the ice edge, the space and time evolution of the pack ice, as well as ice types, thickness distribution, strength and state of deformation is crucial to accurate sea ice forecasting.

We propose to improve the representation of sea ice mechanics in general circulation models and Earth system models in order to advance the short-term (days to months), high- resolution (tens of meters to kilometers) predictive capabilities of these models. We will pursue this objective by expanding, evaluating, and applying a new sea ice model, which describes sea ice mechanics based on the elastic-decohesive rheology. In contrast to the isotropic, viscous-plastic rheology of most current sea ice models, the elastic-decohesive rheology explicitly represents the presence and direction of sea ice deformation features. The specific focus of the present proposal is on ‘understanding’ rather than ‘prediction’; that is, we aim to demonstrate that the elastic-decohesive model can better describe the underlying mechanisms responsible for regional sea ice variation and change through comparison with satellite and in-situ observations of sea ice deformation. If successful, however, this project has the potential to deliver models with much improved predictive capability of high-resolution, coupled ocean, wave, atmosphere, and sea ice processes.

The proposed work falls into three categories: model development, model evaluation, and the application of the new model. Model development focuses on the connection of an ice thickness distribution to the decohesive rheology and implementation of the rotation of the lead direction as leads are advected with the flow of the pack ice. Further, we propose a mechanism to account for the increasing strength of ice growing in leads. These developments are crucial to correctly represent the anisotropy of the ice strength, which impacts the large-scale flow of the ice cover and its local deformation. Model evaluation begins with a comparison to observations of ice concentration, thickness, and drift but focuses on the deformation rate and lead statistics. A variety of airborne and satellite derived data sets will be used. Finally, we will apply the new model in a suite of sensitivity simulations to test its dependency on spatial resolution as well as its ability to simulate break-up events of the sea ice cover in late winter and spring. Timing and spatial extent of ice beak-up are relevant to both climate processes and operational decision making.

This proposal is being submitted in response to the NOAA call, Climate Variability and Predictability (CVP) FY 2015: Understanding Arctic Sea Ice Mechanisms and Predictability. The proposed model supports NOAA’s mission to assess current and future states of the climate system in order to identify potential impacts and inform science, service, and stewardship decisions. The model specifically aims to advance the CVP program goal of understanding Pan-Arctic sea ice interactions by better describing mechanisms involved in prediction of regional sea ice variation and change.

Program (s): Climate Variability & Predictability

Competition: Decadal Climate Variability and Predictability

Award Number: NA20OAR4310407, GC20-208 | View Publications on Google Scholar

The global mean sea level is rising as the result of ocean warming and melting of terrestrial glaciers and ice sheets. Regional sea level changes can deviate significantly from the global average change, which is largely the result of the spatial redistribution of heat and freshwater by ocean circulation. The latter is often simplified by a zonal integral of meridional velocities known as the Meridional Overturning Circulation (MOC). The MOC-modulated meridional divergence of heat and freshwater drives the large-scale steric (due to density changes) sea level changes. Near the coast these changes provide background conditions that, when combined with the effect of tides and with synoptic sea level fluctuations due to the variable atmospheric pressure and winds, can result in nuisance flooding and storm surge events that oftentimes affect densely populated urban areas. The goal of this proposal is to establish the relationships between the MOC and the gyre-scale and coastal sea level changes throughout the Atlantic Ocean basin, and identify the key mechanisms responsible for these relationships. The main focus will be at understanding how the large-scale sea level patterns influenced by the MOC may affect coastal sea level in the Atlantic Ocean and the U.S. East Coast, in particular. The proposed research will utilize a suite of observational data collected by Atlantic MOC observing arrays (e.g., RAPID/MOCHA/WBTS, MOVE, SAMBA, OSNAP, and a combination of altimetry and hydrography, etc.); satellite measurements of sea surface height (SSH), temperature (SST), salinity (SSS), and winds; hydrographic data (Argo, XBT, CTD); coastal tide gauges; and atmospheric re-analyses. Statistical techniques (e.g. Empirical Orthogonal Functions (EOFs), including joint and complex EOFs, wavelet and cross-wavelet transforms, wavelet coherence, etc.) will be used to identify the leading modes of variability, how these modes evolve in space and time, and lag-lead relationships between the modes and other variables. The established relationships will be investigated in more detail using an eddy-permitting Estimating Circulation and Climate of the Ocean Version 5 (ECCOv5) state estimate. The use ECCOv5 will allow budget closures and better process understanding. In addition, the ECCOv5 adjoint sensitivities will be used to quantitatively evaluate the causal mechanisms of regional (including coastal) sea level and heat content variability, as well as changes in the Atlantic MOC. The proposed research directly addresses the second priority area of the FY19 CVP Decadal Climate Variability and Predictability call: “Investigation of the relationship between the Atlantic Meridional Overturning Circulation (AMOC) and global and regional sea level (historical, current, and/or future), with a focus on understanding sea level extremes and coastal impacts in the United States, for the improved understanding of the ocean-climate system”. It also addresses NOAA’s goals by (i) contributing to understanding and predicting changes in climate, weather, oceans and coasts; and (ii) collecting and analyzing information critical for conservation and management of coastal and marine ecosystems and resources.

Program (s): Climate Variability and Predictability

Competition: Improved Understanding of Tropical Pacific Processes, Biases, and Climatology

Award Number: GC14-244 | View Publications on Google Scholar

Among the biases in the tropical Pacific that are common in the climate models, two stand out. One is the excessive cold-tongue in the mean state---the pool of the cold water that is normally in the eastern tropical Pacific extends too far to the west. The other is the underestimate of the asymmetry of El Nino-Southern Oscillation—the fact that El Nino and La Nina are more or less a mirror image of each other in the models while they are not so in the observations. The importance of the tropical Pacific sea surface temperature in affecting climate variability on a range of time-scales over the continental U.S and the world at large demands our attention to the causes and removal of these two common biases in climate models.

The proposal attempts to delineate the relationship between these two common biases in the stateof- the-art climate models and isolate the root causes of them. Toward this objective, we will conduct focused data analysis as well as numerical experiments with climate models of varying complexity. The hypothesis the proposed project sets out to test is that these two tropical biases are linked. More specifically, we suspect that an excessive cold-tongue in the mean climatological state renders the two phases of ENSO more symmetric, possibly through its impact on the stability of the ENSO system, while a more symmetric ENSO results in less nonlinear heating to the cold-tongue which in turn contributes to the development of an excessive coldtongue. Moreover, we suspect that these two biases may be the symptoms of a single structural inadequacy in the models: a weak dynamical coupling between the atmosphere and ocean.

To test our hypothesis and pin down the physical processes responsible for the aforementioned biases, we will

(1) Capitalize on the greater range of variability among the CMIP5 models than CMIP3 models in their simulations of the tropical Pacific climate to examine the relationship between the zonal extent of the cold-tongue and ENSO asymmetry in the models.

(2) Conduct coupled experiments with two models of intermediate complexity to delineate the mechanisms by which an excessive cold-tongue affect the asymmetry of ENSO.

(3) Conduct forced ocean GCM experiments with surface forcing from models with different level of biases as well as from observations to quantify the feedback from ENSO events.

(4) Evaluate the precipitation-wind-SST relationship in the corresponding AMIP runs in conjunction with the coupled runs to fully evaluate the coupling strength as well as to diagnose causes for the initial error

(5) Conduct experiments with a fully coupled GCM (the NCAR Community Climate System Model-version 4) as well as experiments with its atmospheric and oceanic components.

The proposed project utilizes data analysis and models of varying complexity to achieve a deeper understanding of the causes of two prominent biases in the tropical Pacific and thereby helps to improve the simulations and predictions of the tropical Pacific climate—a key source for climate variability and predictability in the earth’s climate system-- by our state-of-the-art climate models. Thus, the proposed project is highly relevant to the objectives and priorities of the Earth System Science Program of NOAA.

Program (s): Climate Variability and Predictability

Competition:

Award Number: | View Publications on Google Scholar

To be posted

Program (s): Climate Variability and Predictability

Competition:

Award Number: | View Publications on Google Scholar

The US National Science and Technology Council's Joint Subcommittee on Ocean Science and Technology (JSOST) has identified the behavior of the Atlantic Meridional Overturning Circulation (AMOC) and its relationship to abrupt climate change as an important research priority. General circulation models (GCMs) play a crucial role in this endeavor. We propose to explore the interplay of deterministic and stochastic processes and their role in the predictability of the AMOC and Atlantic Climate, including the identification of systematic compensating model errors, in AMOC simulations in IPCC GCMs. We shall use novel dynamically-based statistical methods at multiple timescales, both in the frequency and in the temporal domains. We propose to apply Linear Inverse Modeling (LIM) to the output of each GCM to summarize nonlocal interactions between temperature and salinity resolved at the annual timescale, while estimating frequency-dependent transfer functions (transfer function analysis, or TFA) between these variables. Using these methods in combination aids us in separating forced-response multivariate phenomena from processes whose transient behaviors are coupled but operate on different timescales. Phase information from TFA and Fluctuation- Dissipation theory will be combined with LIM results to estimate the subscale forcing, both atmospheric and oceanic, required to maintain the AMOC as represented in each model. The proposed study will localize the sensitive regions affecting the AMOC in each model, identify sources of that sensitivity, diagnose compensating model errors, and allow comparison of results among the different models.

Relevance NOAA's Next-Generation Strategic Plan and to ESS: This project satisfies all foci for the ESS Program, Priority 2, and is directly relevant to NOAA's Next-Generation Strategic Plan. We shall study AMOC variability and compensating errors in IPCC models by estimating state-dependent tendencies as a function of variable and geographical location. Identifying rapidly varying processes (i.e., weather) that interact and/or maintain slower decadal variability (i.e., climate) will elucidate model dependence on those mechanisms. Thus, we may investigate the role of multi-scale interactions in both enhancing and destroying decadal predictability and improve the credibility of those models on which an informed society depends.

Program (s): Climate Variability and Predictability

Competition:

Award Number: | View Publications on Google Scholar

Prediction of sea level rise from understanding and modeling of glacial and land-based ice sheet melt is difficult at best, yet of critical importance for future climate prediction. Antarctic glacial melt is particularly difficult, leading to the Antarctic's contribution to sea level rise being downplayed during IPCC assessment IV. Numerous observation and modeling studies cite the ocean as providing the source of heat for the recently observed acceleration of the Antarctic melt rate. That melt is concentrated in the West Antarctic, at the coastal margin of the Amundsen/Bellingshausen Seas (ABS). We approach this project with 17 years of gridded ocean data adjacent to the West Antarctic Peninsula (WAP) upstream of the West Antarctic Ice Sheet (WAIS) primary drainage basin. These data show that the ocean heat content on the WAP shelf (QWAP) has been rising steadily since the early 1990s, and dramatically since the 1950s, qualitatively consistent with the dramatic increase in the observed glacial melt, and with the required ocean heat. This warm water, Upper Circumpolar Deep Water (UCDW) is available for melting ice in the WAP and WAIS. We desire to determine the ultimate source of this increased ocean heat content, to estimate future warming associated with the source. 

The world oceans have been absorbing heat at their surface from the warming atmosphere, and some of that heat has penetrated to depth, leading to excess ocean heat content (Qexcess); multiple studies argue that the observed Qexcess is due to absorption of anthropogenic heat. Some of this heat will reach, or is already within, southward flowing deep currents transporting it to the Antarctic Circumpolar Current (ACC), where it warms the warm deep subsurface water it already transports. The ACC transports this warmed water to the ABS shelves (the only shelves in the Antarctic where the ACC flows along the shelf-break) fueling accelerated glacier melt. The goal of this project is to assess what fraction of QWAP warming is from this Qexcess and via extrapolation, how much of this Qexcess would still be available for accelerated glacial melt in the ABS, even if there is a reduction or elimination of global warming. 

The analyses will involve assessment of global historical data, beginning as a natural extension to the series of studies that have analyzed historical data to show that the ocean heat content has risen since the 1950s. We will use previously developed objectively analyzed 5° gridded composites to deal with the sparse data deeper than 300 m in earlier decades, updated to correct for XBT and Argo float biases. In our case the focus is more on change as a function of time and space, in an effort to track potential paths of heat transfer to the south (and time scales of the transfer). 

We expect to reveal that amount of heat (with uncertainties) that will still be available for glacial melt regardless of changes in the rate of global warming (or even better, as a function of total global warming). In other words, climate change is already committed to accelerated glacial melt from this stored heat — knowing the magnitude and timescale of its delivery is an essential component required in our ability to model the contribution of glaciers and land-based ice sheets to future global sea level rise. Alternative methods for explaining increasing QWAP (e.g., changes in the strength of the polar westerlies) are being investigated elsewhere, but preliminary analysis of the post-1990s data suggest that both mechanisms contribute. 

Program (s): Climate Variability & Predictability

Competition: OAR/CPO/CVP - NWS/OSTI/Modeling Division - Joint Competition to Advance Process Understanding and Representation of Precipitation in Models

Award Number: NA22OAR4310605 | View Publications on Google Scholar

Extratropical cyclones, which make up the mid-latitude storm tracks, are the key driver producing cool season precipitation in the CONUS. Hence, biases and errors in cyclone and storm track prediction, or in the structure of cyclone related precipitation, can give rise to biases in model predicted precipitation. S2S prediction models can have biases either in the mean climate or in the prediction of climate variability. Hence, precipitation biases and errors over the CONUS can arise due to model biases and errors in: 1) the prediction of climate drivers such as ENSO, MJO, polar vortex and the QBO; 2) the mid-latitude teleconnection patterns associated with these climate modes; 3) the response of extratropical cyclones to these large scale teleconnections; and 4) the precipitation structure associated with extratropical cyclones. Preliminary results by the PIs’ groups have shown that S2S model simulations exhibit several of these biases. This project will diagnose extratropical cyclone related precipitation biases, including biases in extreme precipitation, using GFDL’s SPEAR model simulations. We will evaluate the model bias in winter cyclone frequency/intensity using reanalysis data and examine the linkage between cyclone frequency/intensity bias and precipitation bias at the S2S time scale. Precipitation biases will be assessed using rain gauge- and satellite-based precipitation estimates. Model biases will be stratified according to geographical regions, cyclone paths and evolution, as well as cyclone intensity. Biases in model cyclone and precipitation response to the modes of climate variability discussed above will be quantified. We will diagnose the cyclone related synoptic precipitation structural errors in model simulations using observations and identify the key processes causing these structural errors. Model sensitivity studies will be conducted to provide insights on how these biases may be reduced. Sensitivity to model resolution, sea surface temperature biases, biases in tropical forcing, biases in the large-scale circulation, and biases in model dynamics and physics will be assessed using model intervention experiments. The outcome from this work will inform GFDL’s model development team on how to improve cool season precipitation simulation and prediction. On top of that, our results will provide a detail account of the contribution of different kinds of extratropical cyclones to the mean and extreme precipitation over CONUS, as well as how well SPEAR simulates these contributions. Selected analyses will be conducted to examine biases exhibited by CFSv2 and GEFSv12 subseasonal predictions. This project aims to identify and understand model precipitation biases and systematic errors associated with extratropical cyclones and storm tracks, which are the key physical drivers for precipitation over the CONUS during the cool season, through data analysis and global modeling experiments. While the numerical experiments will be conducted using the GFDL SPEAR model, our diagnostic studies will also be conducted on CFSv2 and GEFS simulations to identify biases and systematic errors in these models. Diagnostic tools developed in this project can also be applied to diagnose model errors and biases in future model simulations and predictions, including those of the subseasonal UFS. Our results will provide insights to model developers on the sources of cool season precipitation biases, informing future model development, thus this project is clearly relevant to the competition and to NOAA’s mission.

Program (s): Climate Variability & Predictability

Competition: OAR/CPO/CVP - NWS/OSTI/Modeling Division - Joint Competition to Advance Process Understanding and Representation of Precipitation in Models

Award Number: NA22OAR4310621 | View Publications on Google Scholar

Due to the coupled nature of the earth system, precipitation forecast errors at subseasonal-to-seasonal (S2S; 2 weeks to 2 months) lead times are caused by a combination of errors/biases from the atmosphere, ocean, ice and land across a range of spatial and temporal scales. Unrealistic sensitivity of model convection to its large-scale environment, as well as unrealistic strength of prominent feedbacks (e.g. cloud radiative, wind-evaporation), can lead to the inability to maintain subseasonal tropical convection variability in forecasts. Even if models were able to perfectly simulate tropical fields, errors in the subtropical and midlatitude circulations can further introduce forecasts errors in U.S. precipitation via incorrect teleconnections. This means that identifying the correct combination of model biases that lead to specific precipitation forecast errors is incredibly challenging. Furthermore, these different processes are not always active at any given time (e.g. the Madden-Julian oscillation can be in an inactive state), implying that their associated biases only contribute to forecast errors intermittently. Thus, an additional challenge in understanding model forecast errors at S2S lead times is identifying the intermittent states of the system when these biases are most important. Here, we propose to couple novel explainable artificial intelligence (XAI) techniques with process-oriented diagnostics to identify, understand, and correct via post-processing, state-dependent UFS precipitation forecast errors. The proposed work is organized into three distinct activities that focus on improving UFS forecasts of North American precipitation at S2S lead times. Activity I involves refining and then implementing an XAI framework to identify state-dependent UFS precipitation errors. Activity II revolves around understanding the tropical-extratropical teleconnection processes relevant to the climate states identified by the XAI method. To do this, we will use process-oriented model diagnostics and test our understanding with simplified model experiments. Activity III aims to leverage what we have learned in Activities I and II to develop XAI-derived post-processing corrections to improve UFS precipitation forecasts for specific initialization states.