Sort by: Project | Lead PI

An Objective Seasonal Drought Outlook for the Conterminous United States

Lead PI: Lettenmaier, Dennis (UCLA)

Co-PI: Kingtse Mo (NOAA/CPC)

The goal of this proposal is to develop an objective drought seasonal outlook (ODSO) over the Conterminous United States (CONUS) for agricultural and hydrological users. The ODSO will utilize both statistical and dynamical forecasts, and will include estimates of their uncertainties. The idea is that forecasters can use the ODSO as their ‘first guess’ for the NOAA Drought Outlook. The proposed project will advance CPC/NOAA operational capabilities for prediction of drought by incorporating the results of past NOAA/MAPP-funded (and other) drought research. To achieve our goal, we plan (1) to better understand drought evolution from onset to demise based on observations of precipitation (P) and temperature (Tair), and model- reconstructed soil moisture (SM), snow water equivalent (SWE, where applicable), evapotranspiration (ET) and runoff (RO) using a suite of land surface models (LSMs) that comprise an updated NLDAS system from 1916-near present; (2) to examine forcings associated with drought and to test statistical tools such as ensemble canonical correlation analysis (ECCA) for forecasting of drought indices; (3) to evaluate the ability of an ensemble of global forecast and land surface models (GCM_LSM) to predict drought development by diagnosing sources of model errors, with global forecasts taken from the North American Multi Model Ensemble (NMME) archive, (4) to utilize the ECCA to optimally combine statistical forecasts and error-corrected dynamical forecasts into an ODSO. The proposed work will directly contribute to the MAPP objectives of advancing drought understanding and prediction. The proposed research will also contribute to the NIDIS objectives of (1) improving drought prediction skill and (2) improving drought information systems. The PI and co-PI will continue to contribute to NOAA’s Drought Task Force (DTF) research to better understand the physical mechanisms and advance the ability to predict various aspects of drought including onset, duration and recovery. The work we propose will be readily transferrable to operations, and in fact we propose to implement a prototype system at CPC that will draw from the operational North American Land Data Assimilation System (NLDAS) and will produce information for CPC Drought Outlook forecasters in near real-time. The TRL level of the proposed project is TR5 (System/subsystem/component validation in relevant environment).

Biosphere-atmosphere regulations of droughts assessed using microwave and solar-induced fluorescence observations and improved plant water stress representation

Lead PI: Gentine, Pierre (Columbia University)

Co-PI: Rongqian Yang (NOAA EMC), Michael Ek (NOAA/EMC), Alexandra Konings (Stanford University)

The proposed work will encompass three intertwined approaches: 1) a better representation of plant water stress and soil moisture, by implementing a plant hydraulic model and new stomatal conductance model in the NOAH land model of the CFS operational forecast model; 2) linking the model surface fluxes and vegetation water status to remote sensing observations of surface fluxes based on solar-induced fluorescence (SIF – a proxy for photosynthesis), and plant water stress strategy based on microwave vegetation optical depth (VOD – directly related to vegetation water status), and 3) defining a quantitative assessment of the model improvement through changes in land-atmosphere interaction using an improved causal statistical model that explicitly defines the time scales of the feedback from weekly to interannual, and its impact on droughts.

In part 1, we will implement a plant hydraulic model that directly resolves the transport of water throughout the plant xylem and an improved stomatal conductance model based on optimal carbon gain. This will define a physically-based response of plants to water stress through either stomatal closure (demand break down) under reduced leaf water potential or through xylem cavitation (supply break down).

In part 2, we will constrain this new model and avoid over-parameterization by making use of a recent vegetation water stress index derived from vegetation optical depth (VOD) measurements from the ASMR satellite, as it can directly sense the canopy water status at a spatial scale (~25 to 50km) relevant to weather and land-surface modeling prediction. We will also use novel estimates of surface fluxes based on solar-induced fluorescence (SIF) measurements to further constrain the dynamics of plant transpiration and the new model parameterization.

We will test the implemented models with typical metrics (2m temperature and humidity, precipitation and geopotential heights) but also with a new causal statistical technique based on multivariate Granger causality. We recently showed that such a technique could be used to assess vegetation-atmosphere feedback and to isolate the time scales of the feedback from monthly to interannual in coupled land-(ocean-)atmosphere models. We will extend this analysis to non-stationary/non-periodic time series using temporal wavelets (as opposed to Fourier transform currently used) so that we can further highlight the time dependence of the feedback in the coupled model. This will further help refine the model during droughts.

This work is directly relevant to the MAPP drought competition 1 as well as to NOAA’s long-term climate goals as it “advances the understanding, monitoring, and prediction of drought [with] improved understanding of predictability relevant to drought, [and] model development.” It also uses “new monitoring [...] products”. It will also improve subseasonal to seasonal prediction, which is critical for water resource and ecosystem health forecasts.

Clarifying the influence of the multiscale coupling between land surface, shallow and deep convection, and large-scale circulation on the predictability of summer drought over the US Great Plains

Lead PI: Fu, Rong (UCLA)


Current dynamic seasonal predictions show lower prediction skills over the US Great Plains than over the eastern and western United States (US). This is particularly so over that region during the warm season (May through August). The dynamical seasonal prediction systems have virtually no prediction skill, for the warm season rainfall over this region. This lack is in large part because the predictability provided by ENSO and tropical Atlantic sea surface temperature anomalies (SSTA) becomes weaker in the warm season. Internal atmospheric variability has significant impact on warm season rainfall, but without strong feedbacks, it can only sustain an anomalous drought circulation for a few weeks. Many studies have suggested land surface feedbacks as the main source of the dry memory. However, such feedbacks cannot sustain dry memory for much more than a month.

Extreme droughts are resulted from dry memory lasting from 3 to 8 months. An empirical model based on such drought memory shows higher skills for seasonal prediction of warm season rainfall anomalies than those of the dynamic models over the US Great Plains. Thus, there is more predictability than that shown by the current dynamic prediction system. However, the processes behind such higher empirical predictability are unclear, limiting our ability to tap into it as an additional source of predictability for the warm season drought. This proposal addresses this knowledge gap by clarifying the following questions:

  • How does the interplay between anomalous land surface conditions, shallow and deep convection, diabatic heating, and large-scale anticyclonic circulation influence the onset, duration and demise of the drought on a seasonal scale over the Great Plains?
  • What role do sub-seasonal atmospheric variability play in the onset and demise of the drought memory on a seasonal scale?
  • What are the causes of drought memory in dynamic models such as those of CFSv2, CCSM4, CESM1 being weak and short-lived?

We will use CFSR, MERRA, NARR reanalyses, and NLDAS land surface product to diagnose the large-scale atmospheric and land surface anomalous conditions. We will also use long-term ground based and satellite datasets, such as the NCDC rainfall data, the North American Soil Moisture Database (NASMD) data, the Vegetation Drought Response Index (VegDRI), the Sun Induced Fluorescence (SIF), CloudSat data, to diagnose the anomalous conditions of cloud, rainfall, surface temperature, soil moisture and vegetation. The Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) atmospheric measurements and Oklahoma Mesonet will be used for detailed process studies. We will evaluate the seasonal predictions from the CFSv2, CCSM4 and CESM1 of the North-American Multi-Model Ensemble (NMME) prediction system.

Collaborative Research: Toward operational predictions of persistent drought driven by multi-year La Niña

Lead PI: Okumura, Yuko (University of Texas)

Co-PI: Pedro DiNezio (University of Texas at Austin), Clara Deser (NCAR)

The interannual cooling of the equatorial Pacific associated with La Niña events has been known to cause drought conditions over the southern tier of the US by displacing the subtropical jet and storm track northward. Whereas El Niño usually terminates after one year, La Niña often persists for 2 years or longer, exacerbating its impacts. It is therefore critical to predict the occurrence of such events with sufficient lead time. Our ongoing work funded by the NOAA CPC MAPP Program laid the groundwork for the prediction of La Niña-driven drought. Through “perfect model” experiments using the Community Earth System Model version 1 (CESM1), we showed that the duration of La Niña is predictable with lead times of 18-24 months. Furthermore, our observational analyses indicate that multi-year La Niña events have robust, sustained impacts on cold season precipitation deficits over the southern half of the US. To further advance our research toward operational predictions of persistent drought driven by multi-year La Niña, we propose the following three sets of research activities:

Task 1 (La Niña duration): We will assess the predictability of La Niña duration in the real world using two suites of long-range CESM1 forecasts initialized with observed oceanic conditions every July and November during 1954-2014. July and November are chosen because they correspond to the peak of El Niño and subsequent discharge of heat from the equatorial Pacific, respectively, after which La Niña typically develops. We will analyze the predictability of La Niña duration and the dynamical processes underlying that predictability for all events occurring during 1954-2014.

Task 2 (La Niña teleconnections): We will assess the predictability of atmospheric teleconnections forced by La Niña by quantifying the influence of various factors including internal atmospheric variability using the long-range CESM1 forecasts and two other suites of atmospheric model simulations in which tropical Pacific SSTs are constrained to observed values during 1880-2015. We will also analyze the mechanisms responsible for the evolution of the teleconnections over the course of multi-year La Niña events using a simple atmospheric model, as well as the atmospheric component of CESM1.

Task 3 (La Niña impacts on drought): We will investigate how the cold season precipitation deficits induced by La Niña teleconnections affect drought over the southern US during the course of multi-year La Niña events through analyses of observational data. We will use soil moisture and accumulated precipitation anomalies as indicators of drought. In particular, we will examine how precipitation deficits during the second cold season of multi-year La Niña events increase the severity and duration of drought compared to single-year La Niña events. We will also explore the predictability of La Niña-driven drought with the same suites of atmospheric model simulations used in Task 2.

Persistent drought driven by multi-year La Niña exerts tremendous socioeconomic impacts on the southern tier of the US. The proposed research activities will fill important gaps in our current understanding of La Niña-driven drought by exploring the predictability of three underlying processes: the duration of La Niña (Task 1); atmospheric teleconnections driven by La Niña (Task 2); and impacts of La Niña teleconnections on drought (Task 3). The project will thus represent a major contribution toward the Drought Task Force activities and a stated objective of the MAPP Competition: “advancing drought understanding, monitoring and prediction”, as well as NOAA’s long-term climate goal: “an informed society anticipating and responding to climate and its impacts”.

Developing an automated weekly probabilistic and categorical drought outlook based on U.S. Drought Monitor and ensemble prediction

Lead PI: Luo, Lifeng (Michigan State University)

Co-PI: Youlong Xia (NOAA/EMC), ZENGCHAO HAO (Beijing Normal University, China)

Drought monitoring and prediction are critical components of the NOAA-led National Integrated Drought Information System (NIDIS). The U.S. Drought Monitor (USDM) has played an important role in gathering, synthesizing and disseminating drought information to a range of users and stakeholders. The USDM is reasonably realistic with multiple sources of information; it is simple and easy to understand with five intensity categories; and it is up-to-date with a fixed weekly release. These features have made the USDM very successful and popular among drought users and stakeholders. It is the USDM drought map that policymakers and media use in discussion of drought and in allocating drought relief. However, on the prediction side, there is currently no suitable drought outlook that matches up with the USDM although monthly and seasonal drought outlooks are issued each month by CPC.

Our team has identified five important gaps in NOAA’s drought prediction capability, and this proposed project will develop an automated, weekly probabilistic and categorical drought outlook to fill these gaps. The proposed research builds off existing operational products, such as the USDM, LIS-based NLDAS, CFSv2 and NMME seasonal forecast, and research outcomes such as the weekly ensemble drought prediction system developed at MSU and the statistical modeling framework for producing probabilistic forecast of USDM drought categories from monthly drought indicators. All the model and data products are readily available; the key methods for utilizing these model and data products to produce a probabilistic and categorical drought outlook have been developed and tested in recent research. This project is a natural step towards integration of the them to produce the desired drought outlook that is multiple model- based, objective, probabilistic, categorical, and can be run on a weekly schedule to match exactly the USDM schedule. To achieve this, the project comprises of six well designed tasks from multimodel offline simulation and seasonal forecast with the LIS-based NLDAS framework to the development and evaluation of the ordinal regression model for predicting the USDM drought categories. The combination of dynamical modeling and statistical modeling is the major strength of this project. These tasks are well connected to ensure the success of the project. This proposed drought outlook system can potentially be seamlessly integrated with the current USDM to provide simple, easy-to-understand and up-to-date drought forecast information to users of USDM.

This proposal responds directly to Competition 1 (Advancing drought understanding, monitoring and prediction) of the MAPP program for FY2017. More specifically, the objectives of this project are in line with several priority areas highlighted in the MAPP information sheet. The project will help to advancing drought prediction system and outlooks operated, used, and produced by NOAA that contribute to the Drought Early Warning System effort. It will also contribute to the development of new national-scale monitoring and forecast products that can help integrate the results of research advances into improved information for managers and communities. From a practical perspective, the automated system with improved skill will provide additional assistance to CPC forecasters to improve their drought outlook.

Developing National Soil Moisture Products to Improve Drought Monitoring

Lead PI: Ford, Trent (Southern Illinois University)

Co-PI: Steven Quiring (Ohio State University); Jessica Lucido (USGS)

This research addresses the critical need to enhance the accuracy and precision of national drought monitoring products by integrating new sources of soil moisture data. Accurate monitoring and prediction of drought events, particularly on sub-seasonal to seasonal timescales, is vital for reducing societal vulnerability through improving drought early warning and drought response. Soil moisture is a key source of information that helps to identify the onset and characterize the severity of agricultural, hydrological, and socioeconomic drought. Accordingly, previous results suggest in situ and satellite remotely sensed soil moisture provide valuable information for drought monitoring and prediction (Anderson et al. 2013; Ford et al. 2015a). Recent advances in satellite-based soil moisture remote-sensing (SMAP and SMOS), land- surface modeling (NLDAS-2 and NASA LIS) and integration of in situ observations (National Soil Moisture Network and North American Soil Moisture Database) provide an opportunity to enhance the spatial resolution and accuracy of national soil moisture products in the United States. These soil moisture products are important for drought prediction and drought monitoring because soil moisture is strongly linked to the onset, severity, and demise of drought events (Hunt et al. 2009; Hayes et al. 2011). However, soil moisture data are collected by many agencies and organizations in the United States using a variety of instruments for diverse applications. These data are often distributed and represented in disparate formats, posing significant challenges for reuse. In response to this challenge the President's Climate Action Plan called for the development of a National Drought Resilience Partnership to better manage drought-related risks by linking information with drought preparedness and longer-term resilience strategies (The White House, 2013). One component of the National Drought Resilience Partnership is developing a coordinated national soil moisture network. The National Integrated Drought Information System has funded a series of workshops and a pilot project to advance this goal. The pilot has served as a reference architecture for a Coordinated National Soil Moisture Network (NSMN) and ultimately demonstrated that in situ soil moisture sensor data could be integrated in real-time from a variety of sources and made accessible at a single common endpoint (Lucido et al., in preparation). During the recent NSMN workshop, held in Boulder, CO in June 2016, the top priority identified by the workshop participants from research and operations communities was near-real-time, national soil moisture products that integrate in situ, satellite-derived, and modeled soil moisture. Given the demand and importance of near-real-time, national soil moisture data the goal of this project is to develop a national-scale drought monitoring product integrating multiple, diverse sources of soil moisture information to improve drought monitoring activities (e.g., U.S. Drought Monitor) and drought early warning. This goal will be achieved by addressing three main objectives:

1) Assess the fidelity of various satellite remote sensing- and model-based soil moisture
products using the North American Soil Moisture Database stations as a benchmark.
2) Integrate remote sensing and modeled soil moisture information with in situ measurements
to develop a national-scale, near-real time soil moisture product for drought monitoring.
3) Design and develop a proof-of-concept cyber infrastructure for delivery of the gridded soil
moisture product

This project specifically addresses the MAPP Competition 1 priority area of advancing operational drought monitoring systems to improve the Nation’s capacity to manage drought- related risk. We will be using datasets from the North American Soil Moisture Database, NLDAS- 2, and U.S. Drought Monitor. Our project goals are closely aligned with the mission of the NOAA Drought Task Force (i.e., achieve significant advances in the ability to monitor and predict drought over North America, and progressing drought early warning systems and experimental drought monitoring tools).

Drought Onset and termination Across North America: Mechanisms and predictability

Lead PI: Seager, Richard (Columbia University)

Co-PI: Mingfang Ting (LDEO), Naomi Henderson (LDEO), Dong Eun Lee (LDEO)

Introduction to the problem of drought onset and termination (DO&T). Much advance has been made identifying the mechanisms that cause drought in North America. The tropical oceans are key drivers but internal atmospheric variability often plays a de- termining role while land surface feedbacks can be important. Less clear are the processes that determine DO&T and whether these have similar controls and predictability. However
society always faces questions of when a drought will end or begin?

Rationale for proposed work. To develop understanding of the mechanisms in the at- mosphere, land and ocean that cause DO&T to enable more accurate prediction of DO&T.

Brief summary of proposed work. DO&T will be defined in terms of change in soil moisture over timescales ranging from a season to a year. The observational basis will be soil moisture data from the atmosphere-forced NLDAS-2 land surface models, multiple atmospheric reanalyses, and sea surface temperature (SST) datasets, for the overlapping 1979 to present period. Multiple statistical methods will be deployed to identify continental scale associations within observations between DO&T and changes in atmospheric circula- tion, storm tracks, moisture transports/convergence, evapotranspiration, runoff and SSTs. DO&T mechanisms in key regions of interest (southwest, Texas/Southern Plains, Cornbelt)
will be analyzed. We will identify the observed mechanisms of DO&T for case studies (the 1999-2004 southwest, the 2010/11Texas/Southern Plains, the 2012 Cornbelt droughts). Identification of physical mechanisms will come from interpretation of ensembles of simulations with multiple SST-forced models and the North American Multimodel Ensemble coupled model hindcasts and forecasts. The relative roles of internal atmosphere variability and SST forcing in causing DO&T will be determined as well as whether models are capable of capturing observed mechanisms of DO&T and how predictable DO&T is. For the three case studies we will generate large ensembles of the SST-forced NCAR CAM5.3 model (i) with and (ii) without soil moisture initialized to the observed state from NLDAS-2 prior to DO&T and (ii) a run of the land surface model component with initialization and forced by observed meteorology and simulate the period of DO&T. Comparison of simulations and observations will determine the mechanisms and predictability of DO&T.

Relevance to Competition and NOAA’s long term climate goals. The work addresses MAPP goals “Developing a better understanding of sources of predictability towards improving predictions of drought onset, evolution, and termination on subseasonal to in- terannual timescales” and “the role of atmospheric, oceanic and land processes and coupled interactions in providing predictability for droughts in North America” and “how predictability sources and processes linked to drought are represented in state-of-the-art modeling and prediction systems”. The work contributes to NOAA long term goals of “scientific understanding, monitoring, and prediction of climate and its impacts, to enable effective decisions” within the identified themes of “Research to advance scientific understanding” and “Modeling and prediction” focused on “weather and climate extremes” and “climate impacts on water resources”.

Exploring process and scale dependencies on the predictability and variability of drought in the United States

Lead PI: Barlage, Michael (NCAR/Research Applications Laboratory)

Co-PI: Zong-Liang Yang (University of Texas at Austin), Fei Chen (NCAR), David Gochis (NCAR)

Two major land surface/hydrologic modeling systems currently in U.S. forecast operations are the North American Land Data Assimilation System (NLDAS; Mitchell et al. 2004) at NCEP and the National Water Model (NWM; Gochis et al. 2013) at NWC. Both systems have some commonalities: applications to drought forecasting and monitoring, the use of the Noah-MP land surface model (LSM), and execution over what is effectively a CONUS domain. However, the NWM is based on a single land-hydrology system while NLDAS employs a multi-LSM ensemble approach. The NLDAS operates on a 0.125° (~12km) spatial resolution grid while the NWM runs on a much higher resolution 1km grid.

We hypothesize that although NLDAS and NWM have different missions and end-user communities, leveraging the high-resolution of the NWM with the multi-model ensemble approach of NLDAS adds value to improving US drought prediction. The overarching goal of the proposed work is to enhance coordination and communication between these two modeling systems to advance the monitoring and prediction of drought by (1) determining the spatial scales at which processes such as terrain-influenced snowpack and groundwater are necessary to capture the dominant drought/hydrology signals, (2) quantifying the hydrology prediction uncertainty through an intelligent selection of model process ensemble members, and (3) producing estimates of spatially- and temporally-varying soil and vegetation parameters to be used in the NWM and NLDAS modeling systems.

The project goal will be attained through a series of tasks that utilize the NWM and NLDAS, namely:
1. Creating a hydrologic landscape heterogeneity scale to determine where (and potentially
when) increased spatial heterogeneity is needed to capture drought-relevant processes;
2. Determining physical processes, and hence ensemble members, that are necessary in both
NWM and NLDAS to improve hydrologic prediction;
3. Leveraging parameter estimation efforts that can be shared between systems.

Project deliverables include:
1. Enhanced understanding of spatial scales and physical processes critical for US drought
monitoring and prediction;
2. Improved NLDAS and NWM systems that can share drought information;
3. A development pathway for both NLDAS and NWM to leverage the advantages of both systems.

Improving the Drought Monitoring Capabilities of Land Surface Models by Integrating Bias-Corrected, Gridded Precipitation Estimates

Lead PI: McRoberts, Brent (Texas A&M University)

Co-PI: Steven Quiring (Ohio State University), Brad Zavodsky (NASA SPoRT), John Nielsen-Gammon (Texas A&M University), Jonathan Case (ENSCO, Inc. / NASA SPORT)

Increasingly, the spending of drought relief money has become more reliant on high-resolution drought assessment, so it is vital we can accurately depict drought on fine spatial scales. An important advancement in our drought assessment capability is the integration of high-resolution drought information into land surface models (LSMs) such as those comprising the North American Land Data Assimilation System (NLDAS). Improvement in the depiction of land use, soil type, and vegetation allows estimation of drought-informative parameters such as soil moisture, evapotranspiration, and streamflow at fine spatial resolutions. However, the accuracy and representativeness of the precipitation data in NLDAS is lagging behind the other information.

We plan on improving precipitation forcing by integrating a radar-based quantitative precipitation estimate (QPE) product in place of the currently operational dataset that uses daily gauge analysis from by the Climate Prediction Center (CPC). Relative to the CPC analysis, gridded QPEs have complete spatial coverage at high resolution using data from radars, satellites, and gauges. A drawback to using the gridded QPEs is the presence of biases related to errors in radar returns. However, we have developed methods for correcting for beam blockage and also mean-field, range-dependent, and two-dimensional biases in a extensively tested three-step algorithm.

We will integrate bias-corrected, gridded QPEs into the NLDAS precipitation forcing dataset to improve the modeling of drought informative variables in LSMs. Testing will optimize the integration bias-corrected NWS QPEs into NLDAS dataset. We will also develop more reliable historical probability distribution functions (PDFs) to improve drought assessment capabilities. We will validate that using bias-corrected, gridded QPEs leads to more accurate estimates of drought-informative variables. The Noah LSM and the NASA Short-term Prediction Research and Transition (SPoRT) Center’s model outputs of soil moisture, water and energy budget variables will be compared to reliable observation-based datasets to validate expected improvements.

This project “Improving the Drought Monitoring Capabilities of Land Surface Models by Integrating Bias-Corrected, Gridded Precipitation Estimates” directly addresses the Modeling, Analysis, Predictions, and Projections (MAPP) competition initiative of advancing drought understanding, monitoring and prediction. The project specifically targets an improvement in drought monitoring by integrating reliable, high-resolution precipitation information into the NLDAS framework. Integration of an improved precipitation dataset into land surface models will advance our ability to predict drought. The project addresses NOAA’s long-term climate goal of aiding “mitigation and adaptation efforts” by providing “sustained, reliable, and timely climate service” and will directly benefit NIDIS, the NOAA Drought Task Force, NLDAS, and the U.S. Drought Monitor.

Representing human-managed influences through thermal product data assimilation in NLDAS: Impacts on the terrestrial water budget and drought estimation

Lead PI: Peters-Lidard, Christa (NASA Goddard)

Co-PI: David Mocko (SAIC at NASA/GSFC), Christopher Hain (NASA/MSFC), Sujay Kumar (NASA/GSFC/HSL), Youlong Xia (NOAA/EMC)

The North American Land Data Assimilation System (NLDAS) has a long successful history of producing products that are used for drought monitoring and numerous other water applications. Recent MAPP-funded efforts led by the PI Peters-Lidard, co-PI Mocko, Co-I Kumar (NASA) and co-I Xia (NCEP/EMC) have demonstrated the utility of remotely-sensed soil moisture, snow, and terrestrial water storage estimates on improving estimates of land- surface conditions and drought characterization within NLDAS. Related work by MSFC Institutional-PI Hain and Collaborator Anderson has shown utility of thermal and vegetation remote sensing products in capturing processes related to human activities, including irrigation water sources/sinks and the effect of burned areas. In the proposed work, we will enable the use of daily vegetation products and the assimilation of thermal remote sensing products in NLDAS. We believe that incorporation of these products into NLDAS LSMs will also produce improved data products for drought monitoring and water resource management that better represent evolving conditions under human-caused and natural changes.

The proposed work will include the following three elements: 1) Using retrospective and operational LAI/GVF products to improve the representation of vegetation within NLDAS. This element will demonstrate the impact of these daily products relative to existing monthly climatologies; 2) Using retrospective and operational thermal-based products to better represent impacts of human-managed agricultural water use and other human-managed influences. All simulations will be performed using the NASA-developed Land Information System (LIS) software framework. Elements #1 and #2 will both also include examination of changes to drought severity/extent through the exploitation of these products; and 3) Performing a thorough evaluation of the current and upgraded NLDAS systems using the NASA-developed Land Verification Toolkit (LVT) for comparison to observations and benchmarking. This element will quantify the improvements to simulated soil moisture, streamflow, and other hydrological fields towards a more realistic and representative drought monitoring system.

The proposed work is highly relevant to objective #3 – Advancing operational drought monitoring systems, with a focus on improving snowpack, streamflow, groundwater, and soil moisture representation; integrating new data sources including remotely-sensed products; accounting for human forcing of droughts; and improving vegetation representation. The capability to retrospectively and operationally use and assimilate thermal and vegetation products using the existing capabilities of LIS for NLDAS will represent a significant advance over the current state-of-the-art in land data assimilation and will directly benefit drought monitoring and assessment, including an expected improvement in the representation of human- managed influences (such as irrigation and drainage). These proposed efforts will also contribute to the MAPP Drought Task Force (DTF) by: 1) providing scientific leadership of the Task Force; 2) linking international efforts on advancing global drought monitoring and prediction; 3) facilitating a drought database and visualization capabilities for drought inter-comparison and benchmarking, and 4) leading narrative reports to connect DTF research to relevant drought events to advance drought understanding through project integration.

Understanding Predictability of Flash Drought over the United States

Lead PI: Wang, Hailan (Science Systems and Applications Inc.)

Co-PI: Randal Koster (NASA Goddard/GMAO)

This proposal aims to investigate causes and predictability of flash droughts over the US. Here flash droughts refer to short-term drought events that have a rapid onset and occur on a time scale of weeks to months, among which the 2012 Great Plains drought is an excellent example. Often occurring during spring and summer, the typical growing season for crops across much of the US, flash droughts can have substantial impacts on agriculture. Recent research by the PIs has shown that leading modes of subseasonal atmospheric circulation variability play a key role in atmospheric circulation and surface meteorological variability on subseasonal time scales and that they have been crucial in the development of recent short-term warm season droughts and heat waves over the US (e.g. the 1988 and 2012 summer droughts). A proper representation of these processes requires a GCM to have a correct simulation of warm season mean state, particularly the Northern Hemisphere (NH) jet streams. Given that many current GCMs, including the NASA GEOS 5 GCM, are deficient in this regard, the PIs have successfully removed much of the mean bias in the NASA GEOS-5 AGCM relative to MERRA-2 reanalysis by applying 6-hourly climatological corrections (relative to MERRA-2) to model basic state variables within the free-running AGCM.

Our proposed work builds on the above development, and has two major thrusts. First, we will investigate causes and physical mechanisms of past flash droughts over the US, and explore potential sources of predictability. This task includes (i) determining key regional processes that led to the development of past US flash droughts and diagnosing their physical origin, and (ii) investigating the separate effects of regional persistent precipitation deficits and heat waves on US flash drought development. Second, we will investigate the predictability of US flash droughts. This includes (i) assessing the prediction skill and predictability of past US flash droughts in the seasonal North American Multi-Model Ensemble (NMME) and subseasonal NMME forecasting systems, (ii) investigating how well the potential sources of predictability identified in the first task and their effects on US flash drought development are represented in the seasonal NMME and subseasonal NMME; and (iii) investigating the impact of model mean bias on forecast skill and predictability estimate of US flash drought, by contrasting the NASA GMAO operational subseasonal forecasts performed using the standard GEOS-5 coupled model (as part of the subseasonal NMME) and a parallel suite of subseasonal forecasts produced with a mean bias-corrected GEOS-5 coupled model.

The link between our work and the NOAA operational drought prediction will be established through our involvement in the NIDIS Drought Prediction and Forecasts Working Group (as a co-Chair and a member). We will communicate with NCEP CPC regularly to ensure that our research findings are applied to improve NOAA operational drought outlook as appropriate. The proposed work directly targets the focus area “Advancing drought understanding, monitoring and prediction” solicited by FY 2017 NOAA MAPP Program. The expected outcome of the proposed work is an improved understanding of the predictability of development of US droughts on subseasonal time scales. It will also contribute to NOAA’s long-term goal of climate adaptation and mitigation through “Improved scientific understanding of the changing climate system and its impacts”.

Understanding the sources of US drought predictability using seasonal reforecasts of sixty years (1958-2017) initialized with multiple land analyses

Lead PI: Huang, Bohua (GMU/COLA)

Co-PI: Chul-Su Shin (GMU/COLA), Paul Dirmeyer (GMU/COLA), Arun Kumar (NOAA/CPC)

Based on a set of recently completed ensemble seasonal reforecasts covering 1958-2014 (to be extended to 2017) using the NCEP Climate Forecast System, version 2 (CFSv2) and initialized from observation-based ocean, land and atmospheric states, we propose to evaluate the predictive skill of US droughts over 60 years and to identify the sources of prediction skill in the ocean, land and global climate trends. As far as we know, this will be the first time that a record of this length of seasonal reforecasts would be used for the study of drought, comparable to the Atmospheric Model Intercomparison Project (AMIP) simulations that are widely used in drought mechanism studies. Taking advantage of this long reforecast dataset, together with AMIP simulations by the CFSv2 atmospheric component with large ensemble sizes, we plan to conduct the following studies of US drought predictions: (1) We will examine whether there are differences in US drought predictioskill in the 1960s-1970s, 1980s-1990s and 2000s-present, associated with different phases of the Pacific decadal oscillation. For this purpose, we will use an advanced statistical method to extract the predictable patterns of the US precipitation from ensemble predictions and examine the connection of these patterns to the ocean, land and atmospheric forcing factors during each of these three periods. We will also examine the mechanisms and model prediction skills of selected major drought events during each of these three periods. Sensitivity experiments will be conducted to identify the potential contributions of specific SST anomalies on the particular events. (2) We will examine the contributions of the observation-based land initial states to the seasonal predictions of the US precipitation and drought events. For this purpose, we will conduct land initialization experiments in which climatological soil moisture will be used to initialize a selected group of drought events. The comparison of these experiments with the reforecasts will isolate the effects of the initial land signals. Furthermore, we will conduct a set of reforecasts using an alternative set of model-based land initial states and specified persisted anomalies to assess the effects of the uncertainty of the land initial states. (3) Using the 60-year reforecasts and the AMIP runs, we will examine the global change trends in evapotranspiration and its potential influence on US droughts by comparing the variability of the temperature, evapotranspiration and precipitation between earlier and later periods (e.g., 1960s-1970s vs. 2000s-present) associated with observed CO2 forcing and initialized or specified boundary conditions. This proposal addresses two of the principal objectives in the MAPP drought project, i.e., developing a better understanding of sources of predictability toward improving predictions of drought onset, evolution, and termination on subseasonal to interannual timescales, and understanding the role of the temperature and evapotranspiration in affecting droughts. The Center for Ocean-Land-Atmosphere Studies and the US Climate Prediction Center will work closely to conduct this collaborative research.

Advancing the Effectiveness and Efficiency of GLDAS Assimilation of JPSS Land Data Products for NCEP NWP and Drought Monitoring Operations

Lead PI: Zhan, Xiwu (NOAA/NESDIS/STAR)

Co-PI: Yin, Jifu (UMD-ESSIC/NOAA-NESDIS-STAR) Zheng, Weizhong (NOAA-NWS-NCEP) Dong, Jiarui (NOAA-NWS-NCEP) Xia, Youlong (NOAA-NWS-NCEP)

For MAPP program Competition 1: Advancing Earth System Data Assimilation, we
propose to advance the effectiveness and efficiency of assimilation of JPSS land data products
through the Global Land Data Assimilation System (GLDAS) for NCEP NWP and drought
monitoring operations. Satellite remote sensing land surface data products have been generated
and made available for various applications in the past decades. The data products of soil
moisture, land surface temperature, albedo, vegetation indices/green vegetation fraction, surface
type are currently operationally generated from NOAA JPSS VIIRS or GCOM-W1/AMSR2
satellite sensors for use in NOAA NCEP numerical weather, climate and hydrological
predictions. Several algorithms have also been introduced to assimilate these data products into
land surface models to improve NCEP weather, climate and hydrological forecasts. However,
none of those JPSS land data products and none of those data assimilation algorithms were used
in NCEP operational NWP models, drought monitoring system, or national water model. The
situation might have been caused by the low efficiency and effectiveness of the currently used
data assimilation algorithms. Based on the findings of the land data assimilation research
community in the past decades, we propose to: 1) implement the dual-pass data assimilation
algorithm by Yang et al (2007) into GLDAS of NCEP without the need to use the so-called
“bias-correction” in the conventional land data assimilation algorithms; 2) inter-compare the
efficiency and effectiveness of the dual-pass data assimilation algorithm with conventional
algorithm (i.e., the EnKF using CDF-matching for bias correction); 3) evaluate the impact of the
dual-pass algorithm in NCEP NWP model simulations and forecasts; 4) examine the drought
monitoring capability using soil moisture profile output from GLDAS with or without
assimilating the satellite land data products; 5) streamline the dual-pass data assimilation
algorithm in GLDAS to enable a straightforward transition of the algorithm to NCEP operation
environment and finally deliver the improved drought monitoring products to NCEP.
Through advancing CPO’s capability of “modeling and prediction” and meeting NOAA’s
long-term climate goal, this proposed project attempts to enhance the effectiveness and
efficiency of assimilation of NOAA JPSS data products in NCEP’s modeling and prediction
systems of GLDAS, Global Forecast System (GFS), Climate Forecast System (CFS) and drought
monitoring operations. This attempt meets CPO’s research objective of focusing on climate
intelligence and climate resilience as stated in the MAPP FY18 information sheet.

Near-real time data assimilation for land vegetation and carbon cycle using JPSS space-based observations and in-situ data

Lead PI: Zeng, Ning (University of Maryland, College Park)

Co-PI: Kleist, Daryl (NOAA/NCEP/EMC) Kalnay, Eugenia (University of Maryland College Park) Kogan, Felix (NESDIS/STAR) Zhan, Xiwu (Jerry) (NESDIS/STAR) Ek, Michael (NCEP/EMC) Dong, Jiarui (NCEP/MC) Zheng, Weizhong (NCEP/EMC)

Satellite remote sensing has been providing observations of the terrestrial biosphere for decades.
For instance, the NDVI index since the 1980s using NOAA’s AVHRR, followed by NASA’s
MODIS, and more recently the JPSS VIIRS instrument onboard the Suomi NPP (SNPP) satellite
that provides a suite of high quality land-surface observations. There is an opportunity to
significantly enhance the utility of the JPSS observations by bringing models and other datasets
in a coherent and comprehensive framework. This proposal will apply a land-vegetation-carbon
Data Assimilation (DA) system to satellite and in-situ observations, with the aim of providing an
assimilated data-model ‘best-analysis’ monitoring product of vegetation and carbon cycle.

This system includes the data assimilation core Local Ensemble Transform Kalman Filter
(LETKF), an advanced DA system that has been successfully applied in a variety of atmosphere,
ocean, and land applications including at NCEP and ECMWF. This assimilator will be applied to
two land-vegetation-carbon models: (1) VEGAS, a UMD-led model widely used in carbon cycle
and dynamic vegetation community, (2) NCEP/EMC’s operational land model Noah with its
dynamic vegetation model version Noah-MP. Weighing observation errors vs. model errors, the
DA system gives an optimized estimate of the state variables such as vegetation biomass, soil
carbon, and leaf area index (LAI). Key model parameters such as light use efficiency (LUE),
respiration dependence on temperature Q10, and phenological cold/drought tolerance are treated
as augmented state variables to be assimilated side-by-side. Gross Primary Productivity (GPP)
and surface fluxes of heat, H2O and CO2 will also be a product of this assimilation.

Our assimilation will include two strands of effort: (1) a historical 37-year period using a
NESDIS/STAR dataset based on harmonized operational global NOAA/AVHRR and 7-year
SNPP/VIIRS data, (2) a shorter current near-real time (NRT) period using the JPSS EDR land
product from VIIRS. In the historical strand, we will build on several datasets/experiences with a
broader suite of datasets, including AVHRR/VIIRS vegetation indices (VIs) and phenology,
FLUXNET carbon and water fluxes across ~200 sites around the world, and the novel remotely
sensed Solar-Induced chlorophyll Fluorescence (SIF) from the Orbiting Carbon Observatory 2
(OCO-2). In the NRT monitoring product, we will focus on the VIIRS vegetation indices
NDVI/EVI and OCO-2 SIF. The historical strand will provide a longer and historical background
that is useful for understanding interannual-decadal variability, as well as validate the system and
establish the best methodology, while the latter will provide an NRT monitoring product of
vegetation health, ecosystem dynamics, fire, agriculture and the global carbon cycle.

The proposal targets at MAPP Competition 1: Advancing Earth System Data Assimilation and
its Objective 2, by developing a new monitoring product on vegetation and carbon cycle. The
deliverable will be a monitoring product of vegetation dynamics and terrestrial carbon cycle at
0.5°×0.5° spatial resolution and hourly time step, provided as a best estimate analysis of blended
satellite and in-situ observations with predictive model. Underlying is a vegetation data
assimilation system that has the potential to provide real-time analysis to initialize vegetation
state for Earth system model prediction, thus contributing to NOAA’s long-term climate and
Earth system modeling goal.

Towards an evolutionary data assimilation system: the value of JPSS Land data in drought monitoring

Lead PI: Moradkhani, Hamid (University of Alabama)

Co-PI: Xiwu Zhan (NOAA-NESDIS)

There is growing concern with evidence that droughts have been intensified due to climate variation and with ongoing land development driven by population growth. This has correspondingly aggravated water scarcity, which threatens the long-term sustainability of water resources. To mitigate the drought vulnerability, an effective drought monitoring system is critical for decision makers. Estimation of key environmental variables such as soil moisture, and evapotranspiration are of paramount importance due to their strong influence on many water resources applications including agricultural production which control the behavior of the climate system. Reliable simulation of land surface properties is highly dependent on the initial and boundary conditions, quality of forcing data, accuracy of measured or calibrated model parameters, and proper estimation of prognostic and diagnostic variables. These can be addressed through Data Assimilation (DA) as a means of merging observations (satellite or in-situ) with model outputs. DA methods based on sequential Bayesian estimation seem better able to take advantage of the temporal organization and structure of information, so that better compliance of the model outputs with observations can be achieved.

This proposal responds to competition 1: Advancing Earth System Data Assimilation-objective 2 by proposing a 3-year collaborative project conducted by investigators from Portland State University and NOAA-NESDIS Center for Satellite Applications and Research. The proposed project aligns well with the NOAA’s long‐term climate goal as described in NOAA CPO’s strategy to “address challenges in the area of Weather and climate extremes, and effectively coordinate across these components through the development and deployment of end‐to-end research‐based integrated information systems that address needs of high societal relevance”. The proposed investigation will be conducted by employing Noah-Multiparameterization (Noah-MP) land surface model and the ensemble data assimilation is implemented by means of a novel sequential Bayesian method, the evolutionary particle filter (PF). An optimal assimilation method is needed to maximize information content from observations and model simulations. PF data assimilation has shown to have strong theoretical foundation providing full probabilistic characterization of prognostic variables which are robust and not prone to violation of mass conservation, and Gaussian assumption of noises in the observations. In this proposal, we plan to utilize available global satellite data products of soil moisture (i.e. JPSS/GCOM-W1/AMSR2 and other products available from Soil Moisture Operational Product System (SMOPS)). In addition, other land surface variables (i.e. JPSS surface type, land surface temperature, albedo, and vegetation indices) will be used as needed during the assimilation process.

We determine seasonal and regional systematic and random errors in model forcing using the best
available resources and reduce systematic errors in the proposed assimilation framework to enhance
monitoring capability. The study enhances the use of remotely sensed satellite soil moisture,
evapotranspiration data as complementary sources of observation to improve prediction of prognostic
and diagnostic hydrologic variables.

In this project, we will use the USDM, the USDA’s disaster declaration, and the drought economic
loss as references to assess the open loop and DA drought monitoring skill. In addition, drought
characteristics such as mean duration, areal extent, total magnitude, intensity, and severity– area–
duration curves will be quantified and verified for the recorded droughts across the CONUS.


1315 East-West Highway Suite 100
Silver Spring, MD 20910


Americans’ health, security and economic wellbeing are tied to climate and weather. Every day, we see communities grappling with environmental challenges due to unusual or extreme events related to climate and weather.