Search

A workshop at MIT discussed the importance of data assimilation and Earth system reanalysis in improving climate models. The workshop emphasized the need to bring models and observations into consistency to reduce uncertainties in climate projections. Recommendations were made for actions and funding in areas such as exploiting Earth observations, developing infrastructure for data assimilation and machine learning, and performing Earth system reanalysis. The document highlights the significance of using observations to improve models and calls for collaboration between different scientific communities to achieve these goals. The recommendations aim to enhance the quality and usability of climate information. [Extracted from the article]

We present an improved approach to generating moist baroclinically unstable background states for f-plane-channel simulations via potential vorticity (PV) inversion. Previous studies specified PV distributions with constant values in the troposphere and the stratosphere, but this produces unrealistic static-stability profiles that decrease sharply with height in the troposphere. Adding moisture to such environments can yield unrealistically large values of convective available potential energy (CAPE) even for reasonable relative humidity (RH) distributions. In our modified approach, we specify a PV distribution that increases with height in the troposphere and the stratosphere, yielding background states with more realistic values of static stability and CAPE. This modification produces environments that are better suited for representing moist processes, namely, deep convection, in idealized extratropical-cyclone simulations. Also, we present a method for introducing moisture that preserves a specified RH distribution while maintaining hydrostatic balance. Our approach allows for a large degree of control over the initial conditions, as background states with different jet strengths and shapes, average temperatures, moisture contents, or horizontal shears can easily be obtained without changing the underlying PV formula and inadvertently producing unreasonable values of static stability or CAPE. We demonstrate the characteristics of idealized extratropical cyclones developing in our background states by adding localized perturbations that represent an upper-level trough passing over a low-level frontal zone. In particular, we illustrate the impacts of horizontal shear, moisture, and grid spacing on baroclinic-wave development. [ABSTRACT FROM AUTHOR]

Subseasonal-to-seasonal (S2S) precipitation prediction in boreal spring and summer months, which contains a significant number of high-signal events, is scientifically challenging and prediction skill has remained poor for years. Tibetan Plateau (TP) spring observed surface temperatures show a lag correlation with summer precipitation in several remote regions, but current global land-atmosphere coupled models are unable to represent this behavior due to significant errors in producing observed TP surface temperatures. To address these issues, the Global Energy and Water Exchanges (GEWEX) program launched the "Impact of Initialized Land Temperature and Snowpack on Subseasonal-to-Seasonal Prediction"(LS4P) initiative as a community effort to test the impact of land temperature in high-mountain regions on S2S prediction by climate models: more than 40 institutions worldwide are participating in this project. After using an innovative new land state initialization approach based on observed surface 2-m temperature over the TP in the LS4P experiment, results from a multimodel ensemble provide evidence for a causal relationship in the observed association between the Plateau spring land temperature and summer precipitation over several regions across the world through teleconnections. The influence is underscored by an out-of-phase oscillation between the TP and Rocky Mountain surface temperatures. This study reveals for the first time that high-mountain land temperature could be a substantial source of S2S precipitation predictability, and its effect is probably as large as ocean surface temperature over global "hotspot"regions identified here; the ensemble means in some "hotspots"produce more than 40% of the observed anomalies. This LS4P approach should stimulate more follow-on explorations.

It is known that the southwest vortex (SWV) is an important weather system that may induce severe weather. The southward deviation of an SWV track forecasted by the Global Assimilation and Prediction System of the China Meteorological Administration (CMA-GFS) is systematically diagnosed in this study. The southward shift of the SWV is directly attributed to the deviation of the steering flow caused by the weak forecast of the upper-level trough. According to the diagnosis of potential tendency, the underestimation of the initial vorticity advection forecasted by CMA-GFS dominates the weak development of the upper-level trough. The underestimation of the vorticity advection is eventually sourced to the weak geostrophic wind caused by the weak initial meridional and zonal gradients of the midlevel height in front of the trough. The assimilation process on the initial field of the CMA-GFS acts a negative effect on forecasting this SWV track. It weakens the π field at midmodel level, resulting in the weak midlevel height gradient in front of the trough. A verified numerical experiment initialized by a more reasonable field is carried out and the southward shift of the SWV is obviously modified. This study suggests that a reasonable analysis field is crucial for the accurate forecast of the SWV track. Significance Statement: The important impact of initial field deviation in key regions on the forecast in the late period is highlighted. A systematic diagnosis process for identifying and addressing forecast issues on SWV track is proposed. This research provides a comprehensive approach for diagnosing the forecast deviation associated with SWV track. [ABSTRACT FROM AUTHOR]

Realistic initialization of the land surface is important to produce accurate NWP forecasts. Therefore, making use of available observations is essential when estimating the surface state. In this work, sequential land surface data assimilation of soil variables is replaced with an offline cycling method. To obtain the best possible initial state for the lower boundary of the NWP system, the land surface model is rerun between forecasts with an analyzed atmospheric forcing. We found a relative reduction of 2-m temperature root-mean-square errors and mean errors of 6% and 12%, respectively, and 4.5% and 11% for 2-m specific humidity. During a convective event, the system was able to produce useful (fractions skill score greater than the uniform forecast) forecasts [above 30 mm (12 h)−1] down to a 100-km length scale where the reference failed to do so below 200 km. The different precipitation forcing caused differences in soil moisture fields that persisted for several weeks and consequently impacted the surface fluxes of heat and moisture and the forecasts of screen level parameters. The experiments also indicate diurnal- and weather-dependent variations of the forecast errors that give valuable insight on the role of initial land surface conditions and the land–atmosphere interactions in southern Scandinavia. [ABSTRACT FROM AUTHOR]

The outputs of the Chinese Academy of Sciences (CAS) Flexible Global Ocean-Atmosphere-Land System (FGOALS-f3-L) model for the decadal climate prediction project (DCPP) of the Coupled Model Intercomparison Project Phase 6 (CMIP6) are described in this paper. The FGOALS-f3-L was initialized through the upgraded, weakly coupled data assimilation scheme, referred to as EnOI-IAU, which assimilates observational anomalies of sea surface temperature (SST) and upper-level (0–1000-m) ocean temperature and salinity profiles into the coupled model. Then, nine ensemble members of 10-year hindcast/forecast experiments were conducted for each initial year over the period of 1960–2021, based on initial conditions produced by three initialization experiments. The hindcast and forecast experiments follow the experiment designs of the Component-A and Component-B of the DCPP, respectively. The decadal prediction output datasets contain a total of 44 monthly mean atmospheric and oceanic variables. The preliminary evaluation indicates that the hindcast experiments show significant predictive skill for the interannual variations of SST in the north Pacific and multi-year variations of SST in the subtropical Pacific and the southern Indian Ocean. [ABSTRACT FROM AUTHOR]

This study assesses the forecast skill of the Canadian Seasonal to Interannual Prediction System (CanSIPS), version 2, in predicting Arctic sea ice extent on both the pan-Arctic and regional scales. In addition, the forecast skill is compared to that of CanSIPS, version 1. Overall, there is a net increase of forecast skill when considering detrended data due to the changes made in the development of CanSIPSv2. The most notable improvements are for forecasts of late summer and autumn target months that have been initialized in the months of April and May that, in previous studies, have been associated with the spring predictability barrier. By comparison of the skills of CanSIPSv1 and CanSIPSv2 to that of an intermediate version of CanSIPS, CanSIPSv1b, we can attribute skill differences between CanSIPSv1 and CanSIPSv2 to two main sources. First, an improved initialization procedure for sea ice initial conditions markedly improves forecast skill on the pan-Arctic scale as well as regionally in the central Arctic, Laptev Sea, Sea of Okhotsk, and Barents Sea. This conclusion is further supported by analysis of the predictive skill of the sea ice volume initialization field. Second, the change in model combination from CanSIPSv1 to CanSIPSv2 (exchanging the constituent CanCM3 model for GEM-NEMO) improves forecast skill in the Bering, Kara, Chukchi, Beaufort, East Siberian, Barents, and the Greenland–Iceland–Norwegian (GIN) Seas. In Hudson and Baffin Bay, as well as the Labrador Sea, there is limited and unsystematic improvement in forecasts of CanSIPSv2 as compared to CanSIPSv1. [ABSTRACT FROM AUTHOR]

Thresholds of soil moisture exist below which the atmosphere becomes hypersensitive to land surface drying, inducing thermal feedbacks that can exacerbate heatwaves. Realistic representation of threshold transitions in forecast models could improve extreme heat predictability and understanding of the role of land–atmosphere coupling. This study evaluates the performance of several forecast models from the Subseasonal Experiment (SubX) and several prototype versions of the Unified Forecast System (UFS) in their representation of threshold transitions by validation against reanalysis data. A metric of skill (true skill score) is applied to soil moisture breakpoint values, which mark the transition to heatwave hypersensitivity for drying soils. Forecast models have poor skill at being initialized on the correct side of the breakpoint, but show improvement when normalized to account for deficiencies in their soil moisture climatologies. Regionally, models performed best in the U.S. Northwest and worst in the Southwest. They effectively capture the tendency of western regions to spend more summer days in the hypersensitive regime than the eastern United States. Models represent well extreme heat as a consequence of atmospheric initial state for the first week of the forecast, but struggle to represent the soil moisture feedback regime. Forecast models generally perform better at extreme heat prediction when they are already dry and in the hypersensitive regime, even when erroneously so, implying that errors or biases exist in model parameterizations. Nevertheless, composite analysis shows encouraging model performance of the "hit" category, suggesting that an improvement in soil moisture initialization could further improve extreme heat forecast skill. [ABSTRACT FROM AUTHOR]

Realistic ocean initial conditions are essential for coupled hurricane forecasts. This study focuses on the impact of assimilating high-resolution ocean observations for initialization of the Modular Ocean Model (MOM6) in a coupled configuration with the Hurricane Analysis and Forecast System (HAFS). Based on the Joint Effort for Data Assimilation Integration (JEDI) framework, numerical experiments were performed for the Hurricane Isaias (2020) case, a category-1 hurricane, with use of underwater glider datasets and satellite observations. Assimilation of ocean glider data together with satellite observations provides opportunity to further advance understanding of ocean conditions and air–sea interactions in coupled model initialization and hurricane forecast systems. This comprehensive data assimilation approach has led to a more accurate prediction of the salinity-induced barrier layer thickness that suppresses vertical mixing and sea surface temperature cooling during the storm. Increased barrier layer thickness enhances ocean enthalpy flux into the lower atmosphere and potentially increases tropical cyclone intensity. Assimilation of satellite observations demonstrates improvement in Hurricane Isaias's intensity forecast. Assimilating glider observations with broad spatial and temporal coverage along Isaias's track in addition to satellite observations further increase Isaias's intensity forecast. Overall, this case study demonstrates the importance of assimilating comprehensive marine observations to a more robust ocean and hurricane forecast under a unified JEDI–HAFS hurricane forecast system. Significance Statement: This is the first comprehensive study of marine observations' impact on hurricane forecast using marine JEDI. This study found that assimilating satellite observations increases upper-ocean stratification during the prestorm period of Isaias. Assimilating preprocessed observations from six gliders increases salinity-induced upper ocean barrier layer thickness, which reduces sea surface temperature cooling and increases enthalpy flux during the storm. This mechanism eventually enhances hurricane intensity forecast. Overall, this study demonstrates a positive impact of assimilating comprehensive marine observations to a successful ocean and hurricane forecast under a unified JEDI–HAFS hurricane forecast system. [ABSTRACT FROM AUTHOR]

An incremental analysis update (IAU) scheme is successfully implemented into a WRF/WRFDA-based hourly cycling data assimilation system with the goal to reduce the imbalance introduced by the high-frequency intermittent data assimilation, especially when radar data are included. With the application of IAU, the analysis increment is smoothly introduced into the model integration over a time window centered at the analysis time. As in digital filter initialization (DFI), the IAU scheme is able to limit large shocks in the early part of a model forecast. Compared to DFI, IAU does better in hydrometeor spinup and produces more continuous precipitation forecasts from cycle to cycle. The run with IAU is shown to improve the precipitation forecast skills (10+% for CSI scores) compared to the regular cycling forecasts without IAU. The data assimilation system with IAU is also able to accept more observations due to balanced first-guess fields. Comparable results are obtained in IAU tests when the time-varying weights are used versus constant weights. Because of its better property, the IAU with the time-varying weights is implemented in the operational system. [ABSTRACT FROM AUTHOR]

Sahel summer rainfall has undergone persistent drought from the 1970s to 1980s, causing severe human life and economic losses. Many studies pointed out that the decadal variations of Sahel rainfall are mainly modulated by low-frequency sea surface temperature (SST) variations in different ocean basins. However, how this modulation contributes to the decadal prediction skill of Sahel rainfall remains unknown. This study provided an affirmative response using the decadal hindcasts initialized by a dimensional-reduced projection four-dimensional variational (DRP-4DVar) data assimilation method to incorporate only ocean analysis data into the gridpoint version 2 of the Flexible Global Ocean–Atmosphere–Land System Model (FGOALS-g2). The hindcasts reveal the benefits of the DRP-4DVar approach for improving the Sahel rainfall decadal prediction skill measured by the anomaly correlation coefficient (ACC), root-mean-square error, ACC difference, and mean square skill score. The decadal variations of SSTs in the Atlantic, Mediterranean Sea, Indian Ocean, and Pacific as well as correct representations of the associated Sahel rainfall–SST relationships are well predicted, thus leading to skillful predictions of Sahel rainfall. In particular, the initialization of SSTs in the Atlantic and Mediterranean Sea plays a more important role in skillful Sahel rainfall predictions than in the other basins. The prediction skill of Sahel rainfall by the FGOALS-g2 prediction system is significantly higher than those by most phase 5 and 6 of the Coupled Model Intercomparison Project (CMIP5&6) prediction systems initialized only with ocean analysis data. This result is likely attributed to a more accurate relationship between Sahel rainfall and SST by the FGOALS-g2 prediction system than by the CMIP5&6 prediction systems. Significance Statement: Previous studies have shown limited success in predicting Sahel rainfall. By using an advanced coupled data assimilation method constrained only by ocean observational data, we achieve a high decadal prediction skill for Sahel rainfall. The successful prediction is attributed to accurately predicted decadal variations of sea surface temperatures in the Atlantic, Mediterranean Sea, Indian Ocean, and Pacific as well as their relationships with Sahel rainfall. Our results can provide references for future decadal predictions of Sahel rainfall and motivate the need to evaluate the contributions of the initialization of the ocean versus the other climate components (e.g., atmosphere or land) to Sahel rainfall predictions. [ABSTRACT FROM AUTHOR]

Self-supervised learning (SSL) has achieved remarkable performance in pre-training the models that can be further used in downstream tasks via fine-tuning. However, these self-supervised models may not capture meaningful semantic information since the images belonging to the same class are often regarded as negative pairs in the contrastive loss. Consequently, the images of the same class are often located far away from each other in the learned feature space, which would inevitably hamper the fine-tuning process. To address this issue, we seek to explicitly enhance the semantic relation among instances on the targeted downstream task and provide a better initialization for the subsequent fine-tuning. To this end, we propose a Contrastive Initialization (COIN) method that breaks the standard fine-tuning pipeline by introducing an extra class-aware initialization stage before fine-tuning. Specifically, we exploit a supervised contrastive loss to increase inter-class discrepancy and intra-class compactness of features on the target dataset. In this way, self-supervised models can be easily trained to discriminate instances of different classes during the final fine-tuning stage. Extensive experiments show that, with the enriched semantics, our COIN significantly outperforms existing methods without introducing extra training cost and sets new state-of-the-arts on multiple downstream tasks. For example, compared with the baseline method, our COIN improves the accuracy by 5% on ImageNet-20 and 2.57% on CIFAR100, respectively. [ABSTRACT FROM AUTHOR]

How do ocean initial states impact historical and future climate projections in Earth system models? To answer this question, we use the 50-member Canadian Earth System Model (CanESM2) large ensemble, in which individual ensemble members are initialized using a combination of different oceanic initial states and atmospheric microperturbations. We show that global ocean heat content anomalies associated with the different ocean initial states, particularly differences in deep ocean heat content due to ocean drift, persist from initialization at year 1950 through the end of the simulations at year 2100. We also find that these anomalies most readily impact surface climate over the Southern Ocean. Differences in ocean initial states affect Southern Ocean surface climate because persistent deep ocean temperature anomalies upwell along sloping isopycnal surfaces that delineate neighboring branches of the upper and lower cells of the global meridional overturning circulation. As a result, up to a quarter of the ensemble variance in Southern Ocean turbulent heat fluxes, heat uptake, and surface temperature trends can be traced to variance in the ocean initial state, notably deep ocean temperature differences of order 0.1 K due to model drift. Such a discernible impact of varying ocean initial conditions on ensemble variance over the Southern Ocean is evident throughout the full 150 simulation years of the ensemble, even though upper ocean temperature anomalies due to varying ocean initial conditions rapidly dissipate over the first two decades of model integration over much of the rest of the globe. [ABSTRACT FROM AUTHOR]

• We comprehensively describe the current state of the individual-based forest landscape and disturbance model iLand after 12 years of ongoing model development. • We highlight key considerations for potential model uses, such as model initialization and evaluation, as well as species parameterization. • We present spatial coverage and research questions addressed in previous model applications and discuss limitations as well as potential avenues for future model development. Forest ecosystems are changing rapidly, and landscape-level processes such as disturbance and dispersal are key drivers of change. Consequently, forest landscape models are important tools for studying forest trajectories under changing environmental conditions and their impacts on ecosystem service provisioning. Here, we synthesize 12 years of development and application of the individual-based forest landscape and disturbance model iLand. Specifically, we describe the fundamental model logic and give an overview of model components introduced over the years. Additionally, we outline how to initialize, evaluate and parameterize the model for new applications. iLand is a process-based forest landscape model that simulates forest dynamics at the level of individual trees. It accounts for continuous processes (tree growth, mortality, and regeneration) as well as discontinuous disturbances (wind, wildfire, and biotic agents) and forest management. Simulations span multiple spatial and temporal scales, from individual trees to landscapes of 105 hectares, and from hourly disturbance dynamics to centuries of forest development. Environmental conditions are represented by daily climate data and high-resolution soil information. The model was designed for flexibly addressing a wide range of research questions, features a rich graphical user interface and comprehensive scripting support. The model is open source and comes with extensive online model documentation. iLand has hitherto been applied in 50 peer-reviewed simulation studies across three continents. Applications primarily focused on the effects of climate change, disturbances and forest management on forest dynamics, ecosystem service provisioning and forest biodiversity. Future model development could address the representation of belowground processes, biotic interactions, and landscape dynamics beyond forest ecosystems. We conclude that process-based simulation of landscape-scale forest dynamics at the level of individual trees has proven a valuable approach of forest landscape modeling. [ABSTRACT FROM AUTHOR]

A dynamical vortex initialization (DVI) scheme is implemented on unstructured meshes for the global model MPAS for typhoon forecasts. The DVI extracts the departure vortex within a specified radius of the vortex center and implants this vortex at the observed vortex location in continuously cycled 1-h integrations of the model. The cycling integration is stopped when either the simulated central sea level pressure or maximum wind speed of the typhoon has reached the value in the best track data, denoted as P-match or V-match, respectively. The DVI may spin up the initial vortex with a more contracting eyewall, but still keeping the same size of the outer vortex. Forecasts for 16 typhoons over the western North Pacific in 2015–20 are investigated. Predictions from the experiments with the 60–15-km variable-resolution MPAS mesh show that both P-match and V-match significantly improve the track forecasts, where V-match mostly requires less cycle runs than P-match. Cycling results with P-match or V-match are also dependent on the choice of physics suites within MPAS. Positive impacts are larger for V-match than P-match using the mesoscale reference physics suite, with significantly improved track forecasts and earlier intensity forecasts. Intensity differences resulting from the DVI have gradually decreased with forecast time, which are closely correlated to the differences in the averaged tropospheric potential vorticity of the inner vortex. The DVI with the 60–15–3-km variable-resolution mesh also works well and improves intensity forecasts. The DVI can also help produce asymmetric structures and spin up inner vortex cores for typhoons near high topography, which leads to improved intensity forecasts. [ABSTRACT FROM AUTHOR]

The Arctic environment is changing, increasing the vulnerability of local communities and ecosystems, and impacting its socio-economic landscape. In this context, weather and climate prediction systems can be powerful tools to support strategic planning and decision-making at different time horizons. This article presents several success stories from the H2020 project APPLICATE on how to advance Arctic weather and seasonal climate prediction, synthesizing the key lessons learned throughout the project and providing recommendations for future model and forecast system development. [ABSTRACT FROM AUTHOR]

Effective exploration is one of the critical factors affecting performance in deep reinforcement learning. Agents acquire data to learn the optimal policy through exploration, and if it is not guaranteed, the data quality deteriorates, which leads to performance degradation. This study investigates the effect of initial entropy, which significantly influences exploration, especially in the early learning stage. The results of this study on tasks with discrete action space show that (1) low initial entropy increases the probability of learning failure, (2) the distributions of initial entropy for various tasks are biased towards low values that inhibit exploration, and (3) the initial entropy for discrete action space varies with both the initial weight and task, making it hard to control. We then devise a simple yet powerful learning strategy to deal with these limitations, namely, entropy-aware model initialization. The proposed algorithm aims to provide a model with high initial entropy to a deep reinforcement learning algorithm for effective exploration. Our experiments showed that the devised learning strategy significantly reduces learning failures and enhances performance, stability, and learning speed. [ABSTRACT FROM AUTHOR]

In this study, we investigate the tendencies of gamma parameters for particle size distributions (PSDs) containing snowflake aggregates in orographic, convective, and stratiform clouds, above snowstorms and above rainstorms, in temperatures ranging from 0° to −45°C. We find a strong relationship between μ and Λ but no dependence on temperature. Higher μ are observed during the experiments sampling winter snowstorms, and lower μ are observed during experiments sampling frozen clouds above convective and orographic storms. We find that a gamma function with a μ of −1.25 provides the best average representation of PSD shape and the most accurate representation of PSD moments related to mass and reflectivity. We also provide a lookup table of maximum particle size boundaries that can be used to parameterize incomplete gamma functions with negative μ values. Significance Statement: In many weather models and satellite retrieval algorithms, frozen clouds and precipitation are governed by the same assumptions even though they develop through different growth processes. This paper provides recommendations for snowflake aggregate size distributions that reflect natural conditions, and these recommended assumptions are demonstrated to improve estimates of mass and radar reflectivity. We studied a variety of storms, such as thunderstorms, snow storms, and winter rainstorms, and we found that our model for snowflake aggregates was nearly identical in all observed conditions. [ABSTRACT FROM AUTHOR]

We compare two seemingly different methods of estimating random error statistics (uncertainties) of observations, the three-cornered hat (3CH) method and Desroziers method, and show several examples of estimated uncertainties of COSMIC-2 (C2) radio occultation (RO) observations. The two methods yield similar results, attesting to the validity of both. The small differences provide insight into the sensitivity of the methods to the assumptions and computational details. These estimates of RO error statistics differ considerably from several RO error models used by operational weather forecast centers, suggesting that the impact of RO observations on forecasts can be improved by adjusting the RO error models to agree more closely with the RO error statistics. Both methods show RO uncertainty estimates that vary with latitude. In the troposphere, uncertainties are higher in the tropics than in the subtropics and middle latitudes. In the upper stratosphere–lower mesosphere, we find the reverse, with tropical uncertainties slightly less than in the subtropics and higher latitudes. The uncertainty estimates from the two techniques also show similar variations between a 31-day period during Northern Hemisphere tropical cyclone season (16 August–15 September 2020) and a month near the vernal equinox (April 2021). Finally, we find a relationship between the vertical variation of the C2 estimated uncertainties and atmospheric variability, as measured by the standard deviation of the C2 sample. The convergence of the error estimates and the standard deviations above 40 km indicates a lessening impact of assimilating RO above this level. Significance Statement: Uncertainties of observations are of general interest and their knowledge is important for assimilation in numerical weather prediction models. This paper compares two methods of estimating these uncertainties and shows that they give nearly identical results under certain conditions. The estimation of the COSMIC-2 bending angle uncertainties and how they compare to the assumed bending angle error models in several operational weather centers suggests that there is an opportunity for obtaining improved impact of RO observations in numerical model forecasts. Finally, the relationship between the COSMIC-2 bending angle errors and atmospheric variability provides insight into the sources of RO observational uncertainties. [ABSTRACT FROM AUTHOR]

Keywords: Numerical weather prediction/forecasting; Coupled models; Data assimilation; Ensembles; Numerical analysis/modeling; Model initialization EN Numerical weather prediction/forecasting Coupled models Data assimilation Ensembles Numerical analysis/modeling Model initialization E1421 E1425 5 06/27/22 20220501 NES 220501 NOAA-NASA Coupled Data Assimilation and Modeling Workshop B I What i : b Scientists came together to discuss potential ideas and specific subjects on data assimilation and Earth system modeling. B I When i : b 20 April 2021 B I Where i : b Online The National Aeronautics and Space Administration (NASA) and National Oceanic and Atmospheric Administration (NOAA) have a long history of collaborations on modeling and data assimilation (DA). Count of workshop participants according to their affiliations, NOAA/STI, NOAA/Physical Science Laboratory (PSL), NOAA/NESDIS/Center for Satellite Applications and Research (STAR), NOAA/NCEP/EMC, NOAA/NCEP, NOAA/GFDL, NASA/PO, NASA/Jet Propulsion Laboratory (JPL), NASA/GMAO, and JCSDA. [Extracted from the article]

We investigate the assimilation of nowcasted information into a classical data assimilation cycle. As a reference setup, we employ the assimilation of standard observations such as direct observations of particular variables into a forecasting system. The pure advective movement extrapolation of observations as a simple nowcasting (NWC) is usually much better for the first minutes to hours, until outperformed by numerical weather prediction (NWP) based on data assimilation. Can nowcasted information be used in the data assimilation cycle? We study both an oscillator model and the Lorenz 63 model with assimilation based on the localized ensemble transform Kalman filter (LETKF). We investigate and provide a mathematical framework for the assimilation of nowcasted information, approximated as a local tendency, into the LETKF in each assimilation step. In particular, we derive and discuss adequate observation error and background uncertainty covariance matrices and interpret the assimilation of nowcasted information as assimilation with an H1-type metric in observation space. Further, we show numerical results that prove that nowcasted information in data assimilation has the potential to significantly improve model based forecasting. [ABSTRACT FROM AUTHOR]

Keywords: Climate models; Data assimilation; Land surface model; Model initialization; Numerical weather prediction/forecasting; Optimization EN Climate models Data assimilation Land surface model Model initialization Numerical weather prediction/forecasting Optimization E733 E740 8 04/11/22 20220301 NES 220301 AIMES Land Data Assimilation Working Group Virtual Workshop on "Tackling Technical Challenges... I B What b i : Over 100 participants from the Earth system modeling and numerical weather prediction land data assimilation communities met virtually to discuss technical challenges faced in developing land data assimilations systems, possible solutions and a roadmap for addressing those challenges, and ideas for building a land DA community to facilitate future collaborations and knowledge exchange. Kick-starting a land DA community with a virtual workshop To begin the process of establishing a land DA community and facilitating knowledge exchange between different land DA groups, we held a 3-half-day virtual workshop in June 2021 with over 100 attendees participating each day. The community may need to adopt DA techniques that are more flexible in terms of coupling to the land model and/or adapt land models to better facilitate their interaction with DA systems. [Extracted from the article]

A subseasonal-to-seasonal (S2S) prediction system was recently developed using the GFDL Seamless System for Prediction and Earth System Research (SPEAR) global coupled model. Based on 20-yr hindcast results (2000–19), the boreal wintertime (November–April) Madden–Julian oscillation (MJO) prediction skill is revealed to reach 30 days measured before the anomaly correlation coefficient of the real-time multivariate (RMM) index drops to 0.5. However, when the MJO is partitioned into four distinct propagation patterns, the prediction range extends to 38, 31, and 31 days for the fast-propagating, slow-propagating, and jumping MJO patterns, respectively, but falls to 23 days for the standing MJO. A further improvement of MJO prediction requires attention to the standing MJO given its large gap with its potential predictability (38 days). The slow-propagating MJO detours southward when traversing the Maritime Continent (MC), and confronts the MC prediction barrier in the model, while the fast-propagating MJO moves across the central MC without this prediction barrier. The MJO diversity is modulated by stratospheric quasi-biennial oscillation (QBO): the standing (slow-propagating) MJO coincides with significant westerly (easterly) phases of QBO, partially explaining the contrasting MJO prediction skill between these two QBO phases. The SPEAR model shows its capability, beyond the propagation, in predicting their initiation for different types of MJO along with discrete precursory convection anomalies. The SPEAR model skillfully predicts the observed distinct teleconnections over the North Pacific and North America related to the standing, jumping, and fast-propagating MJO, but not the slow-propagating MJO. These findings highlight the complexities and challenges of incorporating MJO prediction into the operational prediction of meteorological variables. [ABSTRACT FROM AUTHOR]

Since 2005, NOAA has conducted the annual Intensity Forecasting Experiment (IFEX), led by scientists from the Hurricane Research Division at NOAA's Atlantic Oceanographic and Meteorological Laboratory. They partner with NOAA's Aircraft Operations Center, who maintain and operate the WP-3D and Gulfstream IV-SP (G-IV) Hurricane Hunter aircraft, and NCEP's National Hurricane Center and Environmental Modeling Center, who task airborne missions to gather data used by forecasters for analysis and forecasting and for ingest into operational numerical weather prediction models. The goal of IFEX is to improve tropical cyclone (TC) forecasts using an integrated approach of analyzing observations from aircraft, initializing and evaluating forecast models with those observations, and developing new airborne instrumentation and observing strategies targeted at filling observing gaps and maximizing the data's impact in model forecasts. This summary article not only highlights recent IFEX contributions toward improved TC understanding and prediction, but also reflects more broadly on the accomplishments of the program during the 16 years of its existence. It describes how IFEX addresses high-priority forecast challenges, summarizes recent collaborations, describes advancements in observing systems monitoring structure and intensity, as well as in assimilation of aircraft data into operational models, and emphasizes key advances in understanding of TC processes, particularly those that lead to rapid intensification. The article concludes by laying the foundation for the next generation of IFEX as it broadens its scope to all TC hazards, particularly rainfall, storm-surge inundation, and tornadoes, that have gained notoriety during the last few years after several devastating landfalling TCs. [ABSTRACT FROM AUTHOR]

Short-term spinup for strong storms is a known difficulty for the operational Hurricane Weather Research and Forecasting (HWRF) Model after assimilating high-resolution inner-core observations. Our previous study associated this short-term intensity prediction issue with the incompatibility between the HWRF Model and the data assimilation (DA) analysis. While improving physics and resolution of the model was found to be helpful, this study focuses on further improving the intensity predictions through the four-dimensional incremental analysis update (4DIAU). In the traditional 4DIAU, increments are predetermined by subtracting background forecasts from analyses. Such predetermined increments implicitly require linear evolution assumption during the update, which are hardly valid for rapidly evolving hurricanes. To confirm the hypothesis, a corresponding 4D analysis nudging (4DAN) method, which uses online increments is first compared with the 4DIAU in an oscillation model. Then, variants of 4DIAU are proposed to improve its application for nonlinear systems. Next, 4DIAU, 4DAN and their proposed improvements are implemented into the HWRF 4DEnVar DA system and are investigated with Hurricane Patricia (2015). Results from both the oscillation model and HWRF Model show that 1) the predetermined increments in 4DIAU can be detrimental when there are discrepancies between the updated and background forecasts during a nonlinear evolution; 2) 4DAN can improve the performance of incremental update upon 4DIAU, but its improvements are limited by the overfiltering; 3) relocating initial background before the incremental update can improve the corresponding traditional methods; and 4) the feature-relative 4DIAU method improves the incremental update the most and produces the best track and intensity predictions for Patricia among all experiments. [ABSTRACT FROM AUTHOR]

Reduced surface field (RESURF) technology is one of the most common design optimization criteria of silicon-on-insulator (SOI) lateral power devices. This article proposes a RESURF lateral power device design method using a constrained Bayesian optimization (CBO) framework. By casting the task as a constrained optimization problem, the proposed approach can automatically generate a RESURF structure that meets given breakdown voltage (BV) and ON-resistance (${R} _{ \mathrm{\scriptscriptstyle ON}}$) specifications. The optimization scheme is equipped with two features that can improve the design quality and efficiency. First, design thresholds are used as a safety margin to further assist in reaching the target performance. Second, a model initialization sampling scheme is proposed to guide the sample selection efficiently during initialization steps. Numerical results show that the optimized designs can achieve fully depleted conditions when the breakdown occurs at $\text{N}^{+}\text{N}$ junction and in the body while avoiding the p-n junction breakdown. Additionally, they also show that 1) superior designs can be achieved by introducing the design thresholds and 2) model initialization can boost the efficiency of the optimization model. [ABSTRACT FROM AUTHOR]

A three-dimensional ensemble-variational (3DEnVar) data assimilation algorithm has been developed for the high-resolution AROME NWP system. Building on previous work on 3DEnVar for AROME, we describe a formulation of the 3DEnVar that is based on the traditional square-root preconditioning. The localization may be performed either in gridpoint or spectral space, and allows for cross-covariances between surface pressure and the three-dimensional variables. The scheme has capacity for dual resolution, with the ensemble running at a lower 3.2-km spatial resolution than the deterministic AROME running at 1.3 km. This formulation is compatible with the variational bias correction scheme used in AROME. Hybrid covariances are implemented with climatological covariances at 1.3-km resolution being combined with ensemble perturbations that are interpolated to high resolution on the fly in the computation of the gradient. Hybrid 3DEnVar has an increased computational cost compared to 3DVar, which is partly mitigated by the use of dual resolution and the adoption of a flexible convergence criterion in the minimization. To get the full benefit from the ensemble scheme, it is recommended 1) to increase ensemble size from 25 to 50 members and 2) to decrease the localization length scale for the benefit of high-density radar observations. With those changes, the 3DEnVar outperforms the operational AROME-France 3DVar by a significant margin on the first 12 h of forecast range, as evidenced by a 3-month summer experiment. Finally, a case study reports on the improved prediction of heavy rainfall that frequently occurs in the Mediterranean region. [ABSTRACT FROM AUTHOR]

This study investigates the impact of coupled initialization on the extended-range prediction of the Madden–Julian oscillation (MJO). A set of reforecasts using combinations of the oceanic and atmospheric initial conditions produced with coupled and uncoupled data assimilation (DA) are conducted to evaluate the impact of coupling in the different domains, from the perspective of MJO forecasts. The coupled initial conditions are provided by CERA-SAT pilot coupled reanalysis for the satellite era recently produced by ECMWF. We focus on the prediction skill of the MJO using the real-time outgoing longwave radiation (OLR) MJO index in a series of reforecasts. The impact of atmospheric initial conditions produced by coupled DA shows slight benefit for the MJO prediction. However, compared with the operational ocean reanalysis, the ocean initial conditions created by CERA-SAT degrade the MJO prediction skill during the first 2–3 weeks of the reforecast by 1.5%–5.8%. A moist static energy budget analysis revealed that the underestimation of 0.2 K sea surface temperature, 1.4 W m−2 top of the atmosphere downward longwave radiation, and 3.8 W m−2 latent heat flux over the Maritime Continent leads to a small but statistically significant degradation of the MJO forecast skill. The results demonstrate that the MJO is sensitive to ocean initial conditions, and illustrate the value of the extended-range MJO prediction for evaluating the quality of coupled data assimilation, and suggest that future efforts on coupled data assimilation pay special attention to the balance of air–sea interaction processes over the warm pool area, in terms of modeling, observational needs and system design. [ABSTRACT FROM AUTHOR]

The Swedish Armed Forces configuration of the Weather Research and Forecasting (WRF) Model has problems in forecasting low clouds in stably stratified conditions when the ground is covered by snow. Reforecasts for January and February 2018, together with observations from Sodankylä in northern Finland, are analyzed to find the cause. The investigation is done iteratively between the single-column model (SCM), applied at Sodankylä, and the full 3D version. Our experiments show that the forecast error arises due to inadequate initialization of stratocumulus (Sc) clouds in WRF using the ECMWF global model, Integrated Forecasting System (IFS). By including bulk liquid water and bulk ice water content, from IFS in the initial profile, the downwelling longwave radiation increases and prevents the near-surface temperature from dropping abnormally. This, in turn, prevents artificial clouds from forming at the first model level. When no clouds are present in the IFS initial profile, the Sc clouds can be initialized using information from the observed vertical profiles. Generally, initialization of Sc clouds in WRF improves the forecast substantially. [ABSTRACT FROM AUTHOR]

The impact of the observed sea surface temperature (SST) frequency in the model initialization on the prediction of the boreal summer intraseasonal oscillation (BSISO) over the Western North Pacific (WNP) is investigated using the Beijing Climate Center Climate System Model. Three sets of hindcast experiments initialized by the observed monthly, weekly and daily SST data (referred to as the Exp_MSST, Exp_WSST and Exp_DSST, respectively) are conducted with 3-month integration starting from the 1st, 11th, and 21st day of each month in June–August during 2000–2014, respectively. The results show that the useful prediction skill of BSISO index reaches out to about 10 days in the Exp_MSST, and further increases by 1–2 days in the Exp_WSST and Exp_DSST. The skill differences among various hindcast experiments are especially apparent during the forecast time of 6–20 days. Focusing on the strong BSISO cases in this period, the BSISO activity and its related moist static energy (MSE) characteristics over the WNP are further diagnosed. It is found that from the Exp_MSST to the Exp_WSST and Exp_DSST, the enhanced BSISO prediction skill is associated with the more realistic variations of intraseasonal MSE and its tendency. Among the various budget terms that dominate the MSE tendency, the surface latent heat flux and MSE advection are evidently improved, with reduction of mean biases by more than 21% and 10%, respectively. Therefore, the better reproduced MSE variation may contribute to the more skillful BSISO forecast through improving the surface evaporation as well as atmospheric convergence and divergence that related to the BSISO activity. Our findings suggest the necessity of increasing the observed SST frequency (i.e., from monthly to weekly or daily) in the initialization process of coupled models to enhance the actual BSISO predictability, since some current subseasonal forecast operations and researches still use relatively low-frequency SST observations for the model initialization. [ABSTRACT FROM AUTHOR]

We present a coupled retrospective forecast (hindcast) study using the Met Office Global Coupled Model, version 2 (GC2), in which we identify and mitigate causes of initialization shock that lead to rapid error growth in sea ice forecasts. Sea ice state variables and volume budget terms as a function of forecast lead time are evaluated relative to analyses from an uncoupled Met Office ocean–sea ice analysis system [Forecast Ocean Assimilation Model, version 13 (FOAMv13)]. Two sources of initialization shock are highlighted and addressed, both of which are related to effective differences in physics between the analysis system and coupled forecast model. The primary shock to sea ice state variables arises from the use of a salinity-independent freezing temperature for seawater in GC2 as opposed to a salinity-dependent formulation in FOAMv13. A secondary effect arises from differences in the sea ice roughness and hence air–ice drag in the GC2 forecast model compared to the FOAMv13 analysis system. Generalizing from the findings of this study, we suggest that using nonnative analyses as initial conditions for coupled numerical weather prediction (NWP) studies will likely make them prone to initialization shock in some model components, to the detriment of forecast skill. To reduce the undesirable impacts of initialization shock on short-range forecast skill noted in this study we would therefore recommend the use of initial conditions (analyses) physically consistent with the native model components of the coupled forecast model, a native coupled analysis likely being the optimal initialization method. [ABSTRACT FROM AUTHOR]

Ensemble forecasting is a method to faithfully describe initial and model uncertainties in a weather forecasting system. Initial uncertainties are much more important than model uncertainties in the short-range numerical prediction. Currently, initial uncertainties are described by the ensemble transform Kalman filter (ETKF) initial perturbation method in Global and Regional Assimilation and Prediction Enhanced System-Regional Ensemble Prediction System (GRAPES-REPS). However, an initial perturbation distribution similar to the analysis error cannot be yielded in the ETKF method of the GRAPES-REPS. To improve the method, we introduce a regional rescaling factor into the ETKF method (we call it ETKF_R). We also compare the results between the ETKF and ETKF_R methods and further demonstrate how rescaling can affect the initial perturbation characteristics as well as the ensemble forecast skills. The characteristics of the initial ensemble perturbation improve after applying the ETKF_R method. For example, the initial perturbation structures become more reasonable, the perturbations are better able to explain the forecast errors at short lead times, and the lower kinetic energy spectrum as well as perturbation energy at the initial forecast times can lead to a higher growth rate of themselves. Additionally, the ensemble forecast verification results suggest that the ETKF_R method has a better spread-skill relationship, a faster ensemble spread growth rate, and a more reasonable rank histogram distribution than ETKF. Furthermore, the rescaling has only a minor impact on the assessment of the sharpness of probabilistic forecasts. The above results all suggest that ETKF_R can be effectively applied to the operational GRAPES-REPS. [ABSTRACT FROM AUTHOR]

Idealized first-principles models of chemical plants can be inaccurate. An alternative is to fit a Machine Learning (ML) model directly to plant sensor data. We use a structured approach: Each unit within the plant gets represented by one ML model. After fitting the models to the data, the models are connected into a flowsheet-like directed graph. We find that for smaller plants, this approach works well, but for larger plants, the complex dynamics arising from large and nested cycles in the flowsheet lead to instabilities in the solver during model initialization. We show that a high accuracy of the single-unit models is not enough: The gradient can point in unexpected directions, which prevents the solver from converging to the correct stationary state. To address this problem, we present a way to fine-tune ML models such that initialization, even with very simple solvers, becomes robust. • Purely machine learning-based plant simulation can work even for large plants • In the presence of complex cycles, the gradient needs to be constrained • Fine-tuning with the gradient computed after iterative solving enables convergence • After fine-tuning, solving is straightforward regardless of solver. [ABSTRACT FROM AUTHOR]

Effective exploration is one of the critical factors affecting performance in deep reinforcement learning. Agents acquire data to learn the optimal policy through exploration, and if it is not guaranteed, the data quality deteriorates, which leads to performance degradation. This study investigates the effect of initial entropy, which significantly influences exploration, especially in the early learning stage. The results of this study on tasks with discrete action space show that (1) low initial entropy increases the probability of learning failure, (2) the distributions of initial entropy for various tasks are biased towards low values that inhibit exploration, and (3) the initial entropy for discrete action space varies with both the initial weight and task, making it hard to control. We then devise a simple yet powerful learning strategy to deal with these limitations, namely, entropy-aware model initialization. The proposed algorithm aims to provide a model with high initial entropy to a deep reinforcement learning algorithm for effective exploration. Our experiments showed that the devised learning strategy significantly reduces learning failures and enhances performance, stability, and learning speed.

The model initialization with high-resolution SAR wind data provided by the Sentinel-1 mission and its impact on the meteorological model WRF-ARW simulations is discussed. The activity is performed within the Horizon 2020 CEASELESS project, focusing on one of the target areas, the northern Adriatic Sea (northern-central Mediterranean). The Sentinel-1 SAR wind is ingested into LAPS, a numerical system developed at NOAA, specifically designed for data analysis and nowcasting issues, since it has the advantage of being faster and less computational demanding than advanced data assimilation methods. Here, LAPS analyses are used to perform a smarter initialization of the WRF-ARW model simulations than using simply global model fields. The impact of the Sentinel-1 SAR wind on the model simulations is evaluated for twenty cases, ranging through several atmospheric conditions occurring in different seasons of the years 2014–2018. For each case study, a reference WRF-ARW simulation is forced with GFS analysis and forecasts used as initial and boundary conditions, respectively. Additional model runs are initialized with the LAPS analyses, which include the information of Sentinel-1 SAR wind, METAR data and the SEVIRI/MSG (Eumetsat) brightness temperature. A statistical evaluation of the WRF-ARW simulations is performed versus an independent set of surface records, provided by the Friuli Venezia Giulia regional station network (northeastern Italy), and METAR data. The comparison is performed for 10 m wind, 2 m air and dew point temperature. The results show a positive, albeit modest, impact on the WRF model simulations initialized with the LAPS analyses. The initialization with the Sentinel-1 SAR wind show benefits for all surface variables. Finally, a Mediterranean tropical-like cyclone (Medicane), occurred in the Ionian Sea in November 2017, is considered in order to show how the use of Sentinel wind data can contribute to a better analysis and simulation of severe weather episodes in the Mediterranean. The improvement in the simulation of the pressure minimum location is remarkable.

Abstract The next‐generation seasonal prediction system is built as part of the Seamless System for Prediction and EArth System Research (SPEAR) at the Geophysical Fluid Dynamics Laboratory (GFDL) of the National Oceanic and Atmospheric Administration (NOAA). SPEAR is an effort to develop a seamless system for prediction and research across time scales. The ensemble‐based ocean data assimilation (ODA) system is updated for Modular Ocean Model Version 6 (MOM6), the ocean component of SPEAR. Ocean initial conditions for seasonal predictions, as well as an ocean state estimation, are produced by the MOM6 ODA system in coupled SPEAR models. Initial conditions of the atmosphere, land, and sea ice components for seasonal predictions are constructed through additional nudging experiments in the same coupled SPEAR models. A bias correction scheme called ocean tendency adjustment (OTA) is applied to coupled model seasonal predictions to reduce model drift. OTA applies the climatological temperature and salinity increments obtained from ODA as three‐dimensional tendency terms to the MOM6 ocean component of the coupled SPEAR models. Based on preliminary retrospective seasonal forecasts, we demonstrate that OTA reduces model drift—especially sea surface temperature (SST) forecast drift—in coupled model predictions and improves seasonal prediction skill for applications such as El Niño–Southern Oscillation (ENSO).

In an effort to improve radar data assimilation configurations for potential operational implementation, GSI EnKF data assimilation experiments based on the operational system employed by the Center for Analysis and Prediction of Storms (CAPS) real-time Spring Forecast Experiments are performed. These experiments are followed by 6-h forecasts for an MCS on 28-29 May 2017. Configurations examined include data thinning, covariance localization radii and inflation, observation error settings, and data assimilation frequency for radar observations. The results show experiments that assimilate radar observations more frequently (i.e., 5-10 min) are initially better at suppressing spurious convection. However, assimilating observations every 5 min causes spurious convection to become more widespread with time, and modestly degrades forecast skill through the remainder of the forecast window. Ensembles that assimilate more observations with less thinning of data or use a larger horizontal covariance localization radius for radar data predict fewer spurious storms and better predict the location of observed storms. Optimized data thinning and horizontal covariance localization radii have positive impacts on forecast skill during the first forecast hour that are quickly lost due to the growth of forecast error. Forecast skill is less sensitive to the ensemble spread inflation factors and observation errors tested during this study. These results provide guidance toward optimizing the configuration of the GSI EnKF system. Among the DA configurations tested, the one employed by the CAPS Spring Forecast Experiment produces the most skilled forecasts while remaining computationally efficient for real-time use. SIGNIFICANCE STATEMENT: High-resolution ensemble forecasts that can skillfully predict thunderstorms provide an opportunity for warning severe weather further in advance. However, forecast accuracy is dependent upon many factors including the data assimilation system used to create the forecasts. This study optimizes the design of a real-time ensemble forecast and data assimilation system for a severe weather event where a line of thunderstorms produced hail, wind, and tornadoes in the southern united States. The 0-6-h forecasts predict the thunderstorms with moderate skill. Forecast accuracy is most sensitive to how frequently the data assimilation system assimilates radar observations, the degree to which radar observations are thinned, and the covariance localization radius. Results of this study can be used to design future real-time forecast systems for severe weather events. [ABSTRACT FROM AUTHOR]

A scale-dependent localization (SDL) method was formulated and implemented in the Gridpoint Statistical Interpolation (GSI)-based four-dimensional ensemble-variational (4DEnVar) system for NCEP FV3-based Global Forecast System (GFS). SDL applies different localization to different scales of ensemble covariances, while performing a single-step simultaneous assimilation of all available observations. Two SDL variants with (SDL-Cross) and without (SDL-NoCross) considering cross-wave-band covariances were examined. The performance of two- and three-wave-band SDL experiments (W2 and W3, respectively) was evaluated through 1-month cycled data assimilation experiments. SDL improves global forecasts to 5 days over scale-invariant localization including the operationally tuned level-dependent scale-invariant localization (W1-Ope). The W3 SDL-Cross experiment shows more accurate tropical storm–track forecasts at shorter lead times than W1-Ope. Compared to the W2 SDL experiments, the W3 SDL counterparts applying tighter horizontal localization at medium-scale wave band generally show improved global forecasts below 100 hPa, but degraded global forecasts above 50 hPa. While the outperformance of the W3 SDL-NoCross experiment versus the W2 SDL-NoCross experiment below 100 hPa lasts for 5 days, that of the W3 SDL-Cross experiment versus the W2 SDL-Cross experiment lasts for 3 days. Due to local spatial averaging of ensemble covariances that may alleviate sampling error, the SDL-NoCross experiments show slightly better forecasts than the SDL-Cross experiments at shorter lead times. However, the SDL-Cross experiments outperform the SDL-NoCross experiments at longer lead times, likely from retention of more heterogeneity of ensemble covariances and resultant analyses with improved balance. Relative performance of tropical storm–track forecasts in the W2 and W3 SDL experiments are generally consistent with that of global forecasts. [ABSTRACT FROM AUTHOR]

Snow's thermal and radiative properties strongly impact the land surface energy balance and thus the atmosphere above it. Land surface snow information is poorly known in mountainous regions. Few studies have examined the impact of initial land surface snow conditions in high-resolution, convection-permitting numerical weather prediction models during the midlatitude cool season. The extent to which land surface snow influences atmospheric energy transport and subsequent surface meteorological states is tested using a high-resolution (1 km) configuration of the Weather Research and Forecasting (WRF) Model, for both calm conditions and weather characteristic of a warm late March atmospheric river. A set of synthetic but realistic snow states are used as initial conditions for the model runs and the resulting differences are compared. We find that the presence (absence) of snow decreases (increases) 2-m air temperatures by as much as 4 K during both periods, and that the atmosphere responds to snow perturbations through advection of moist static energy from neighboring regions. Snow mass and snow-covered area are both important variables that influence 2-m air temperature. Finally, the meteorological states produced from the WRF experiments are used to force an offline hydrologic model, demonstrating that snowmelt rates can increase/decrease by factor of 2 depending on the initial snow conditions used in the parent weather model. We propose that more realistic representations of land surface snow properties in mesoscale models may be a source of hydrometeorological predictability [ABSTRACT FROM AUTHOR]

The next‐generation seasonal prediction system is built as part of the Seamless System for Prediction and EArth System Research (SPEAR) at the Geophysical Fluid Dynamics Laboratory (GFDL) of the National Oceanic and Atmospheric Administration (NOAA). SPEAR is an effort to develop a seamless system for prediction and research across time scales. The ensemble‐based ocean data assimilation (ODA) system is updated for Modular Ocean Model Version 6 (MOM6), the ocean component of SPEAR. Ocean initial conditions for seasonal predictions, as well as an ocean state estimation, are produced by the MOM6 ODA system in coupled SPEAR models. Initial conditions of the atmosphere, land, and sea ice components for seasonal predictions are constructed through additional nudging experiments in the same coupled SPEAR models. A bias correction scheme called ocean tendency adjustment (OTA) is applied to coupled model seasonal predictions to reduce model drift. OTA applies the climatological temperature and salinity increments obtained from ODA as three‐dimensional tendency terms to the MOM6 ocean component of the coupled SPEAR models. Based on preliminary retrospective seasonal forecasts, we demonstrate that OTA reduces model drift—especially sea surface temperature (SST) forecast drift—in coupled model predictions and improves seasonal prediction skill for applications such as El Niño–Southern Oscillation (ENSO). Plain Language Summary: Dynamic seasonal prediction systems employ global climate models to predict climate variations on monthly to seasonal time scales. A new state‐of‐the‐art seasonal prediction system has been developed at the Geophysical Fluid Dynamics Laboratory (GFDL) of the National Oceanic and Atmospheric Administration (NOAA). The prediction models are based on GFDL's new component models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6). The new seasonal prediction system includes new ways to apply observational information to initialize the model simulations, including an updated data assimilation system for the Modular Ocean Model Version 6 (MOM6). Bias correction is applied to the coupled dynamic models to reduce prediction bias, namely, the gap between model‐simulated and real‐world climatology. Preliminary seasonal prediction experiments demonstrate reduced errors in the predicted climate state globally, as well as improved prediction skill for applications such as El Niño–Southern Oscillation (ENSO). Key Points: The next‐generation SPEAR seasonal prediction system uses the recently developed coupled general circulation model at GFDLThe updated ocean data assimilation system for the MOM6 ocean model produces ocean state estimation and initial conditions for predictionsThe new coupled model and ocean tendency adjustment (OTA) reduce model forecast drift and improve seasonal prediction skill [ABSTRACT FROM AUTHOR]

In the context of forecasting societally impactful Great Plains low-level jets (GPLLJs), the potential added value of satellite soil moisture (SM) data assimilation (DA) is high. GPLLJs are both sensitive to regional soil moisture gradients and frequent drivers of severe weather, including mesoscale convective systems. An untested hypothesis is that SM DA is more effective in forecasts of weakly synoptically forced, or uncoupled GPLLJs, than in forecasts of cyclone-induced coupled GPLLJs. Using the NASA Unified Weather Research and Forecasting (NU-WRF) Model, 75 GPLLJs are simulated at 9-km resolution both with and without NASA Soil Moisture Active Passive SM DA. Differences in modeled SM, surface sensible (SH) and latent heat (LH) fluxes, 2-m temperature (T2), 2-m humidity (Q2), PBL height (PBLH), and 850-hPa wind speed (W850) are quantified for individual jets and jet-type event subsets over the south-central Great Plains, as well as separately for each GPLLJ sector (entrance, core, and exit). At the GPLLJ core, DA-related changes of up to 5.4 kgm-2 in SM can result in T2, Q2, LH, SH, PBLH, and W850 differences of 0.688C, 0.71 g kg-2, 59.9Wm-2, 52.4Wm-2, 240 m, and 4ms-1, respectively. W850 differences focus along the jet axis and tend to increase from south to north. Jet-type differences are most evident at the GPLLJ exit where DA increases and decreases W850 in uncoupled and coupled GPLLJs, respectively. Data assimilation marginally reduces negative wind speed bias for all jets, but the correction is greater for uncoupled GPLLJs, as hypothesized. [ABSTRACT FROM AUTHOR]

Aims: Predictions of medium to long-term changes in soil organic carbon (SOC) stocks involve simulation models with plant carbon (C) inputs estimated from crop yields using crop harvest indices and allometric scaling involving fixed root-to-shoot ratios. The choice of allometric approach and initial distribution of SOC among model pools is crucial for model estimates. Methods: We evaluated the impact of a modified allometric approach for grass leys and initial model set-up using the C-TOOL model. Four European agricultural long-term experiments were considered including three experiments with different frequency of grass leys in the crop rotation and one permanent grassland. Results: The modified allometric function for ley relied on fixed below-ground C input regardless of mineral fertilizer inputs and modified pool initialization involved available site history. Including available, but insufficient, pre-experiment history to adjust the initial set-up of model SOC pools did not improve to the C-TOOL simulations. Changing the allometric approach for ley from fixed shoot-to-root ratios to fixed below-ground C input decreased the soil C input dramatically and improved the C-TOOL simulation of SOC stocks for fertilized treatments in all experiments when combined with standard model set-up. For unfertilized treatments, however, the efficiency of the standard allometric function was superior to the modified one. Conclusions: Our results challenge the prevalent use of allometric approaches based on fixed root-to-shoot in SOC simulation models. The potential of modifying the initial distribution of SOC among model pools by accounting for historic land use remains in the dark. [ABSTRACT FROM AUTHOR]

Initial abstraction (Ia) is a sensitive parameter in hydrological models, and its value directly determines the amount of runoff. Ia, which is influenced by many factors related to antecedent watershed condition (AWC), is difficult to estimate due to lack of observed data. In the Soil Conservation Service curve number (SCS-CN) method, it is often assumed that Ia is 0.2 times the potential maximum retention S. Yet this assumption has frequently been questioned. In this paper, Ia/S and factors potentially influencing Ia were collected from rainfall–runoff events. Soil moisture and evaporation data were extracted from GLDAS-Noah datasets to represent AWC. Based on the driving factors of Ia, identified using the Pearson correlation coefficient and maximal information coefficient, artificial neural network (ANN)-estimated Ia was applied to simulate the selected flood events in the Hydrologic Engineering Center Hydrologic Modeling System (HEC-HMS) model. The results indicated that Ia/S varies over different events and different watersheds. Over 75% of the Ia/S values are less than 0.2 in the two study areas. The driving factors affecting Ia vary over different watersheds, and the antecedent precipitation index appears to be the most influential factor. Flood simulation by the HEC-HMS model using statistical Ia gives the best fitness, whereas applying ANN-estimated Ia outperforms the simulation with median Ia/S. For over 60% of the flood events, ANN-estimated Ia provided better fitness in flood peak and depth, with an average Nash–Sutcliffe efficiency coefficient of 0.76 compared to 0.71 for median Ia/S. The proposed ANN-estimated Ia is physically based and can be applied without calibration, saving time in constructing hydrological models. [ABSTRACT FROM AUTHOR]

The recently launched Geostationary Operational Environmental Satellite "R-series" (GOES-R) satellites carry the Geostationary Lightning Mapper (GLM) that measures from space the total lightning rate in convective storms at high spatial and temporal frequencies. This study assimilates, for the first time, real GLM total lightning data in an ensemble Kalman filter (EnKF) framework. The lightning flash extent density (FED) products at 10-km pixel resolution are assimilated. The capabilities to assimilate GLM FED data are first implemented into the GSI-based EnKF data assimilation (DA) system and tested with a mesoscale convective system (MCS). FED observation operators based on graupel mass or graupel volume are used. The operators are first tuned through sensitivity experiments to determine an optimal multiplying factor to the operator, before being used in FED DA experiments FEDM and FEDV that use the graupel-mass or graupel-volume-based operator, respectively. Their results are compared to a control experiment (CTRL) that does not assimilate any FED data. Overall, both DA experiments outperform CTRL in terms of the analyses and short-term forecasts of FED and composite/3D reflectivity. The assimilation of FED is primarily effective in regions of deep moist convection, which helps improve short-term forecasts of convective threats, including heavy precipitation and lightning. Direct adjustments to graupel mass via observation operator as well as adjustments to other model state variables through flow-dependent ensemble cross covariance within EnKF are shown to work together to generate model-consistent analyses and overall improved forecasts. [ABSTRACT FROM AUTHOR]

High-quality, high-resolution, hourly unbiased surface (2 m) temperature analyses are needed for many applications, including training and validation of statistical postprocessing applications. These temperature analyses are often generated through data assimilation procedures, whereby a background short-range gridded forecast is adjusted to newly available observations. Even with frequent updates to newly available observations, surface-temperature analysis errors and biases can be comparatively large relative to errors and biases of midtropospheric variables, especially over land, despite more near-surface in situ observations. Larger near-surface errors may have several causes, including biased background forecasts and the spatial heterogeneity of surface temperatures that results from subgrid-scale surface, vegetation, land-use, and terrain variations. Are biased raw background forecasts the predominant cause of surface temperature analysis errors? Part I of this two-part series describes a simple benchmark for evaluating the error characteristics of short-term (1 h) raw model background surface temperature forecasts. For stations with a relatively complete time series of data, it is possible to generate an hourly, diurnally, and seasonally dependent observation climatology at a station. The deviation of the current hour's temperature observation with respect to this hour's and Julian day's climatology is added to the climatology for the next hour. For contiguous U.S. stations in July 2015, the station benchmark was lower in error than interpolated 1-h high-resolution numerical predictions of surface temperature from NOAA's High-Resolution Rapid Refresh (HRRR) system, although not including full postprocessing. For August 2018, 1-h HRRR forecasts were much improved when tested against the station benchmark. [ABSTRACT FROM AUTHOR]

There is a longstanding challenge in numerical weather and climate prediction to accurately model tropical wave variability, including convectively coupled equatorial waves (CCEWs) and the Madden–Julian oscillation. For subseasonal prediction, the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS) has been shown to be superior to the NOAA Global Forecast System (GFS) in simulating tropical variability, suggesting that the ECMWF model is better at simulating the interaction between cumulus convection and the large-scale tropical circulation. In this study, we experiment with the cumulus convection scheme of the ECMWF IFS in a research version of the GFS to understand which aspects of the IFS cumulus convection scheme outperform those of the GFS convection scheme in the tropics. We show that the IFS cumulus convection scheme produces significantly different tropical moisture and temperature tendency profiles from those simulated by the GFS convection scheme when it is coupled with other physics schemes in the GFS physics package. We show that a consistent treatment of the interaction between parameterized convective plumes in the GFS planetary boundary layer (PBL) and the IFS convection scheme is required for the GFS to replicate the tropical temperature and moisture profiles simulated by the IFS model. The GFS model with the IFS convection scheme, and the consistent treatment between the convection and PBL schemes, produces much more organized convection in the tropics, and generates tropical waves that propagate more coherently than the GFS in its default configuration due to better simulated interaction between low-level convergence and precipitation. [ABSTRACT FROM AUTHOR]

A new method for modeling the lowest model level vertical motion is described and validated. Instead of smoothing terrain heights, the new method calculates the terrain gradient on a high-resolution grid and averages the gradient values around a gridpoint location. In essence, the method provides a way to achieve some of the impact of very steep terrain on the flow without the computational overhead associated with the very high grid resolution needed to fully resolve complex terrain. The more accurate depiction of the terrain gradient leads to an increase in orographic vertical motion and causes rainfall to occur more often over the windward-facing mountain slopes, consistent with observations. Model results are compared with rain gauge data during the month of January 2016 as well as radar data from a case study on 9 March 2012. When implemented in the Weather Research and Forecasting (WRF) Model over the island of Oahu and compared with the current WRF method, the model precipitation forecast skill is improved. The new method produces more precipitation over the island during January 2016, which is closer to the observed value. On 9 March 2012, the new method clearly focuses the precipitation over the Ko'olau Mountains, reducing the number of false alarm forecasts by nearly one-half. Although the changes to model precipitation skill were small, they were generally positive. [ABSTRACT FROM AUTHOR]

The advances in communication technologies have made it possible to gather road condition information from moving vehicles in real time. However, data quality must be assessed and its effects on the road weather forecasts analyzed before using the new data as input in forecasting systems. Road surface temperature forecasts assimilating mobile observations in the initialization were verified in this study. In addition to using measured values directly, different statistical corrections were applied to the mobile observations before using them in the road weather model. The verification results are compared to a control run without surface temperature measurements and to a control run that utilized interpolated values from surrounding road weather stations. Simulations were done for the period 12 October 2017–30 April 2018 for stationary road weather station points in southern Finland. Road surface temperature observations from the stations were used in the forecast verification. According to the results, the mobile observations improved the accuracy of road surface temperature forecasts when compared to the first control run. The statistical correction methods had a positive effect on forecast accuracy during the winter, but the effect varied during spring when the daily temperature variation was strong. In the winter season, the forecasts based on the interpolated road surface temperature values and the forecasts utilizing mobile observations with statistical correction had comparable accuracy. However, the tested area has high road weather station density and not much elevation variation, so results might have been different in more varying terrain. [ABSTRACT FROM AUTHOR]