Paul Field
UK Met Office
Exeter, United Kingdom
We use convection permitting global aquaplanet simulations to explore the interaction between aerosol and mid-latitude cyclones. Based on model simulations we propose a hypothesis about how midlatitude cyclones will respond to increases in aerosol loading. In this talk, I will introduce the model results and describe how we tested it with a decade of satellite observations and a more focused period coinciding with the Icelandic volcanic eruption in 2014.
Refreshments: 3:15 PM
Xiaolei Zou
Earth System Science Interdisciplinary Center (ESSIC)
University of Maryland
Abstract: The Advanced Technology Microwave Sounder (ATMS) onboard Suomi National Polar orbiter Partnership (S-NPP) satellite combines Advanced Microwave Sounding Unit-A (AMSU-A) and Microwave Humidity Sounder (MHS) onboard NOAA and Meteorological Operational Satellite Program of Europe (MetOP) satellites to simultaneously provide collocated radiance measurements of the atmospheric temperature and moisture profiles under almost all weather conditions except for heavy precipitation. The two lowest frequency ATMS window channels 1-2 (23.8GHz and 31.4 GHz) are the same as AMSU-A channels 1-2 and the other two high-frequency ATMS window channels 17-18 (88.2GHz and 165.5GHz) are similar to MHS window channels 1-2. These four ATMS window channels can be used together for identifying both liquid and ice cloudy radiances. This important feature of ATMS proved to be important for improving the forecast skill of severe weathers populated with clouds (e.g., hurricanes) through satellite microwave radiance assimilation (Zou et al., 2013). Assimilation of microwave radiance data in numerical weather prediction (NWP) models has traditionally been carried out with AMSU-A and MHS data in two separate data streams since the launch of NOAA-15 in 1998. Inspired by the ATMS data assimilation success, a new approach was proposed to combine AMSU-A and MHS radiances into one data stream for their assimilation. It was shown that the spatial collocation between AMSU-A and MHS field of views (FOVs) allows for an improved quality control of MHS data, especially over the conditions where the liquid-phase clouds are dominate. It was found that the quantitative precipitation forecast (QPF) skill associated with landfall hurricanes was significantly improved by the one data stream approach, resulting from a closer fit of analyses to AMSU-A and MHS observations is obtained, especially for AMSU-A surface-sensitive channels (Zou et al., 2017). A shortcoming was also found for S-NPP ATMS whose radiance observations displayed a clear across-track striping noise, which was not found in AMSU-A radiances. Three algorithms were subsequently developed for mitigating the ATMS striping noise for the upper-level sounding channels (Qin et al., 2013), for an operational implementation (Ma and Zou, 2015) and for surface sensitive channels (Zou et al., 2017). Impacts of striping noise mitigation on observation error variances were also quantified for assimilation of destriped ATMS radiance observations.
Refreshments: 10:45 AM
Sisi Chen
Department of Atmospheric and Oceanic Sciences, McGill University
Montreal, Quebec, Canada
Shallow convective clouds are ubiquitous, and warm rain largely contributes to the total annual rainfall, particularly in the tropics. Therefore, understanding the microphysical processes inside these cloud systems becomes important. Classical parcel models often produce narrow droplet size distributions (DSDs) which disagree with observations in cumulus clouds. Since the last century, turbulence have been postulated to explain the effective DSD broadening in early cloud stage.
This work studies the very fundamental process involving droplet condensational and collisional growth to explore the fast warm-rain initiation using the direct numerical simulation (DNS). DNS model can accurately resolve small-scale turbulence and simulates the turbulence impacts on droplets that are tracked in the Lagrangian framework, which is infeasible in other models.
This is the first modeling study that incorporates both droplet condensational process and collisional process into the DNS model and investigates the full droplet growth history in the turbulent environment.
Model results show that condensational growth by itself produces narrow DSD under small-scale turbulence, which is similar to the parcel model results. Results from the simulations that consider pure collision-coalescence process show that small-scale turbulence significantly increases the collision rate between small droplets and thus accelerates the formation of large droplets. In particular, the enhancement is the strongest between similar-sized droplets, which indicates that turbulence effectively broadens the narrow DSD formed by condensational growth. On the other hand, condensational growth considerably brings tiny droplets to 5-10 microns, dynamically shifting the collision rates of those droplets in turbulence. To study how collisional process and condensational process interact under the effect of turbulence, simulation results that consider both condensational and collisional processes will be compared to pure collision-coalescence case. It is shown that the inclusion of condensation significantly changes the behavior of droplet collisions in the turbulence and thus has strong feedback on the DSD broadening. Detailed results and comparison will be presented in the talk.
Refreshments: 3:15 PM
Dale Barker
UK Met Office
The Met Office global and regional NWP applications are centered around the use of the Unified Model (UM) to provide short-range forecasts out to 5-7 days of global and local significant weather. This talk will describe some of the major upgrades implemented or planned during the timeframe of the new Cray XC40 supercomputer (2015 - 2020) beginning with a brief description of the basic NWP configurations and a summary or recent major upgrades e.g. variational bias correction, additional satellite data, etc.
In July 2017, the resolution of the global NWP system at the Met Office was increased to ~10km, with an associated increase to 20km for the global (MOGREPS-G) ensemble. A more significant change is the introduction of hourly-cycling four dimensional variational (4DVar) data assimilation for the km-scale UK model. The relative contributions to forecast skill improvements of hourly-cycling, the use of the 4DVar technique, and improved driving global model will be assessed in this talk.
Looking forward, additional major upgrades are planned in the next 1-2 years including weakly coupled ocean-atmosphere data assimilation, extension of the km-scale MOGREPS-UK ensemble to T+5 days (plus resolution increase from 2.2km to 1.5km), replacement of the current ETKF ensemble system with an ‘Ensemble of 4D Ensemble Vars’. Details of these promising scientific developments will be provided. Finally, a brief summary of plans for the post-UM ‘Exascale Era’ beginning in ~2023 will be outlined.
Note special date and time.
Refreshments: 10:45 AM
Nedjelika Žagar
University of Ljubljana
Ljubljana, Slovenia
Many studies of the forecast error growth focused on the extra-tropical quasi-geostrophic dynamics and often considered the error-free large-scale initial state. In contrast, the operational global numerical weather prediction and ensemble prediction systems are characterized by uncertainties in the initial state at all scales, especially in the tropics. In this seminar the evidence will be discussed about the dominant role of the large-scale error growth early in the forecasts in comparison with the errors cascades from the smaller scales. A new parametric model for the representation of the error growth will be derived. In contrast to the commonly used models, the new model does not involve computation of the time derivatives of the empirical data. The asymptotic error is not a fitting parameter, but it is computed from the model constants.
Simulated forecast errors by the operational ensemble prediction system of the European Centre for Medium-Range Weather Forecasts are decomposed into scales and the new model is applied independently to every zonal wavenumber. A combination of hyperbolic tangent functions in the parametrization of the error growth proves robust to reliably model complex growth dynamics across many scales. The range of useful prediction skill, estimated as a scale where forecast errors exceeds 60% of their asymptotic values is around 7 days on large scales and 2-3 days at 1000 km scale. The new model is easily transformed to the widely used model of Dalcher and Kalnay (1987) to discuss the scale-dependent growth as a sum of two terms, the so-called a and b terms. Their comparison shows that at planetary scales their contributions to the growth in the first 2 days are similar whereas at small scales the b term describes most of a rapid exponential growth of errors towards saturation.
Refreshments: 3:15pm
Xiaolei Zou
Earth System Science Interdisciplinary Center (ESSIC)
University of Maryland
Abstract: The Advanced Technology Microwave Sounder (ATMS) onboard Suomi National Polar orbiter Partnership (S-NPP) satellite combines Advanced Microwave Sounding Unit-A (AMSU-A) and Microwave Humidity Sounder (MHS) onboard NOAA and Meteorological Operational Satellite Program of Europe (MetOP) satellites to simultaneously provide collocated radiance measurements of the atmospheric temperature and moisture profiles under almost all weather conditions except for heavy precipitation. The two lowest frequency ATMS window channels 1-2 (23.8GHz and 31.4 GHz) are the same as AMSU-A channels 1-2 and the other two high-frequency ATMS window channels 17-18 (88.2GHz and 165.5GHz) are similar to MHS window channels 1-2. These four ATMS window channels can be used together for identifying both liquid and ice cloudy radiances. This important feature of ATMS proved to be important for improving the forecast skill of severe weathers populated with clouds (e.g., hurricanes) through satellite microwave radiance assimilation (Zou et al., 2013). Assimilation of microwave radiance data in numerical weather prediction (NWP) models has traditionally been carried out with AMSU-A and MHS data in two separate data streams since the launch of NOAA-15 in 1998. Inspired by the ATMS data assimilation success, a new approach was proposed to combine AMSU-A and MHS radiances into one data stream for their assimilation. It was shown that the spatial collocation between AMSU-A and MHS field of views (FOVs) allows for an improved quality control of MHS data, especially over the conditions where the liquid-phase clouds are dominate. It was found that the quantitative precipitation forecast (QPF) skill associated with landfall hurricanes was significantly improved by the one data stream approach, resulting from a closer fit of analyses to AMSU-A and MHS observations is obtained, especially for AMSU-A surface-sensitive channels (Zou et al., 2017). A shortcoming was also found for S-NPP ATMS whose radiance observations displayed a clear across-track striping noise, which was not found in AMSU-A radiances. Three algorithms were subsequently developed for mitigating the ATMS striping noise for the upper-level sounding channels (Qin et al., 2013), for an operational implementation (Ma and Zou, 2015) and for surface sensitive channels (Zou et al., 2017). Impacts of striping noise mitigation on observation error variances were also quantified for assimilation of destriped ATMS radiance observations.
Refreshments: 10:45 AM
Sisi Chen
Department of Atmospheric and Oceanic Sciences, McGill University
Montreal, Quebec, Canada
Shallow convective clouds are ubiquitous, and warm rain largely contributes to the total annual rainfall, particularly in the tropics. Therefore, understanding the microphysical processes inside these cloud systems becomes important. Classical parcel models often produce narrow droplet size distributions (DSDs) which disagree with observations in cumulus clouds. Since the last century, turbulence have been postulated to explain the effective DSD broadening in early cloud stage.
This work studies the very fundamental process involving droplet condensational and collisional growth to explore the fast warm-rain initiation using the direct numerical simulation (DNS). DNS model can accurately resolve small-scale turbulence and simulates the turbulence impacts on droplets that are tracked in the Lagrangian framework, which is infeasible in other models.
This is the first modeling study that incorporates both droplet condensational process and collisional process into the DNS model and investigates the full droplet growth history in the turbulent environment.
Model results show that condensational growth by itself produces narrow DSD under small-scale turbulence, which is similar to the parcel model results. Results from the simulations that consider pure collision-coalescence process show that small-scale turbulence significantly increases the collision rate between small droplets and thus accelerates the formation of large droplets. In particular, the enhancement is the strongest between similar-sized droplets, which indicates that turbulence effectively broadens the narrow DSD formed by condensational growth. On the other hand, condensational growth considerably brings tiny droplets to 5-10 microns, dynamically shifting the collision rates of those droplets in turbulence. To study how collisional process and condensational process interact under the effect of turbulence, simulation results that consider both condensational and collisional processes will be compared to pure collision-coalescence case. It is shown that the inclusion of condensation significantly changes the behavior of droplet collisions in the turbulence and thus has strong feedback on the DSD broadening. Detailed results and comparison will be presented in the talk.
Refreshments: 3:15 PM
Dale Barker
UK Met Office
The Met Office global and regional NWP applications are centered around the use of the Unified Model (UM) to provide short-range forecasts out to 5-7 days of global and local significant weather. This talk will describe some of the major upgrades implemented or planned during the timeframe of the new Cray XC40 supercomputer (2015 - 2020) beginning with a brief description of the basic NWP configurations and a summary or recent major upgrades e.g. variational bias correction, additional satellite data, etc.
In July 2017, the resolution of the global NWP system at the Met Office was increased to ~10km, with an associated increase to 20km for the global (MOGREPS-G) ensemble. A more significant change is the introduction of hourly-cycling four dimensional variational (4DVar) data assimilation for the km-scale UK model. The relative contributions to forecast skill improvements of hourly-cycling, the use of the 4DVar technique, and improved driving global model will be assessed in this talk.
Looking forward, additional major upgrades are planned in the next 1-2 years including weakly coupled ocean-atmosphere data assimilation, extension of the km-scale MOGREPS-UK ensemble to T+5 days (plus resolution increase from 2.2km to 1.5km), replacement of the current ETKF ensemble system with an ‘Ensemble of 4D Ensemble Vars’. Details of these promising scientific developments will be provided. Finally, a brief summary of plans for the post-UM ‘Exascale Era’ beginning in ~2023 will be outlined.
Note special date and time.
Refreshments: 10:45 AM
Nedjelika Žagar
University of Ljubljana
Ljubljana, Slovenia
Many studies of the forecast error growth focused on the extra-tropical quasi-geostrophic dynamics and often considered the error-free large-scale initial state. In contrast, the operational global numerical weather prediction and ensemble prediction systems are characterized by uncertainties in the initial state at all scales, especially in the tropics. In this seminar the evidence will be discussed about the dominant role of the large-scale error growth early in the forecasts in comparison with the errors cascades from the smaller scales. A new parametric model for the representation of the error growth will be derived. In contrast to the commonly used models, the new model does not involve computation of the time derivatives of the empirical data. The asymptotic error is not a fitting parameter, but it is computed from the model constants.
Simulated forecast errors by the operational ensemble prediction system of the European Centre for Medium-Range Weather Forecasts are decomposed into scales and the new model is applied independently to every zonal wavenumber. A combination of hyperbolic tangent functions in the parametrization of the error growth proves robust to reliably model complex growth dynamics across many scales. The range of useful prediction skill, estimated as a scale where forecast errors exceeds 60% of their asymptotic values is around 7 days on large scales and 2-3 days at 1000 km scale. The new model is easily transformed to the widely used model of Dalcher and Kalnay (1987) to discuss the scale-dependent growth as a sum of two terms, the so-called a and b terms. Their comparison shows that at planetary scales their contributions to the growth in the first 2 days are similar whereas at small scales the b term describes most of a rapid exponential growth of errors towards saturation.
Refreshments: 3:15pm
James Done
NCAR/MMM
As populations increase in hazard-prone regions, the human, cultural and economic costs rise, and will continue to rise in the future. The likely scenario of the weather and climate hazards themselves changing in the future will compound the problem. A transformation of how weather and climate risk is assessed and integrated with risk management practice is needed for society to confront this new era of weather and climate risk. Bringing physics to bear on risk assessment has the potential to transform our understanding of weather and climate risk. Furthermore, physically based risk assessments that are informed by risk management practice are a potentially powerful component of climate resilience. Three recent examples will be presented to illustrate the flow between physically based weather and climate risk assessments and community action.
The first example is the development of a terrain-aware tropical cyclone wind probability assessment at the global scale. In collaboration with a reinsurance broker, an approach to modeling tropical cyclone wind footprints is developed by fitting a parametric wind field model to historical and synthetic cyclone track data, and bringing the winds down to the surface using a 3-dimensional numerical boundary model, accounting for terrain and surface roughness effects. The new wind probability assessments are being used to understand inland wind risk in regions of complex topography, and assess public and private risk management strategies in regions of sparse historical data. The second example explores how the relationship between residential losses and hurricane winds is modified through building codes. Adherence to the Florida building code drives down losses by up to 70%, and the code is cost-effective with a return on investment after 12 years under current climate. The final example explores the role of decadal climate predictions in water resource and flood risk management. The multi-disciplinary UDECIDE (Understanding Decision-Climate Interactions on Decadal Scales) project combines statistical and physical assessments of climate prediction skill with data from interviews with managers to identify intersections at the decadal scale in support of effective management.
Refreshments: 3:15pm
Lotte Bierdel
Ludwig Maximilians University
Munich, Germany
The current literature discussing predictability of atmospheric flow and the nature of the underlying scale interactions considers the problem from two main perspectives. One approach is based on statistical closure models in a homogeneous and isotropic turbulent flow, where the predictability time is determined solely by the background kinetic energy spectrum and not by the underlying dynamical model. An alternative approach is based on results from numerical weather prediction models that suggest that latent heat release associated with deep moist convection is a primary mechanism for small-scale error growth. From this point of view error growth in the atmosphere is an initially localized, highly intermittent phenomenon that expands upscale, leading to a complete loss of predictability on scales below 100 km within a few hours. The error growth process then depends on the underlying dynamics of the respective scale range and the errors in particular have to transition from geostrophically unbalanced to balanced motion while propagating through the mesoscale. In this talk a study will be presented that examines the geostrophic adjustment process as possibly underlying this transition. To that end, an analytical framework for the geostrophic adjustment of an initial pointlike pulse of heat (modeling a convective cloud or an error within the prediction of a cloud) is developed. Spatial and temporal scales of the geostrophic adjustment mechanism are deduced and three characteristics of the solution are shown to be potentially useful for identifying the geostrophic adjustment process in numerical simulations. These three predictions are then tested in the framework of error growth experiments in idealized numerical simulations of a convective cloud field. Three different rotation rates are employed in order to identify the geostrophic adjustment mechanism and allow a quantitative comparison with the predictions of the analytical model. As will be shown, the numerical simulations agree well with the predictions developed from the analytical model. Based on these findings it is suggested that the geostrophic adjustment process governs upscale error growth through the atmospheric mesoscales.
Refreshments: 3:15 PM
Stipo Sentic
New Mexico Tech
Tropical convective organization is the process in which disorganized convection organizes into regions of intense convective activity surrounded by dry, convectively inactive regions. Well known examples are tropical cyclones and the Madden-Julian Oscillation (MJO)—they affect atmospheric energetics, and the MJO affects virtually all weather on our planet. Recent advances in idealized modelling of tropical convection, namely the weak temperature gradient (WTG) approximation, enable us to study convective organization in idealized settings. The WTG approximation parameterizes the effects of the large-scale on local convection, and can be used in idealized sensitivity studies of convection to changes in large-scale convective environment. To model organized convection in the context of the MJO, we used observations from the Dynamics of the Madden-Julian Oscillation (DYNAMO) field campaign to force WTG simulations in a cloud resolving model, and test how well the WTG approximation reproduces variations in convective diagnostics: precipitation rate, stability, moisture content, and large-scale transport (gross moist stability). We find that the WTG approximation reproduces variations in these diagnostics, and relationships between them. The ability of WTG approximation to reproduce important observed diagnostics provides confidence that this is a good strategy for exploring tropical phenomena. An example that I'll talk about is the behavior of convective organization at different SSTs.
(Special Date) Tuesday, 30 May 2017, 3:30 PM
Refreshments: 3:15 PM
NCAR-Foothills Laboratory
3450 Mitchell Lane
Bldg. 2, Small Seminar Room 1001 (Note Location)
Michael Tjernström
Department of Meteorology & Bolin Centre for Climate Research
Stockholm University, Sweden
Arctic climate is ultimately determined by a balance between meridional heat transport into the area, and radiation heat loss at the top of the atmosphere over the same area. Since the net radiation loss is due to small-scale processes parameterized in models, and the meridional heat flux is due to larger scale atmospheric dynamics resolved by the models, the two has usually been studied separately. In this seminar this concept will be called into question.
In an episode during the Arctic Clouds in Summer Experiment (ACSE) in the summer of 2014, warm air from the Siberian mainland flowed in over melting sea-ice in the East-Siberian Sea for over a week. As the ~25 °C warm air flowed over the melting surface, maintained at the melting point, a strong surface inversion formed in which dense fog also formed. This resulted in a positive net longwave radiation while the sensible heat flux was downward. Although solar radiation was attenuated by the fog, this led to an additional 10-20 Wm-2 energy to the surface. This led us to hypothesize a zone from the ice edge where the surface will receive enhanced energy when the atmospheric flow is northward onto the ice.
To test this hypothesis, we analyzed the observation from the entire ACSE expedition. All temperature profiles taken over sea ice were categorized into cases with or without a surface inversion; the inversion cases where further divided into two categories using the humidity profiles. When projecting other observations onto these three classes, many are systematically different. Surface inversion with increasing moisture with height systematically added 10-20 Wm-2 energy to the surface energy budget, indicating that meridional heat flux must be considered together with the small-scale processes caused by the air mass transformation.
Please note the location change.
Thursday, 1 June 2017, 3:30 PM
Refreshments: 3:15 PM
NCAR-Foothills Laboratory
3450 Mitchell Lane
Bldg. 2, Small Seminar Room 1001
Sharon Sessions
Department of Physics, New Mexico Tech
Socorro, New Mexico
Tropical convection is difficult to understand and even more difficult to predict, in part because of the interplay between the convection itself and the large scale circulations. Predictability is possible, however, if the scales of convective disturbances are large enough that they are influenced by voriticity anomalies in the environment. Ooyama, in 1982, discussed this idea in the context of mature tropical cyclones, in a process he refered to as "cooperative intensification". Recently, Raymond et al. (2015) revisted Ooyama's ideas and addressed the question of whether other less extreme types of tropical disturbances could be a response to a nonlinear form of "balanced dynamics". If so, they argued that these types of disturbances would have potential for predictability (and therefore would also be parameterizable). In terms of time scales, disturbances which occur on scales longer than the time to establish balance, are candidates for predictability based on the potential for moist convection to evolve as a balanced response to large scale vorticity anomalies.
In this talk, I'll revisit some of Ooyama's and Raymond's ideas regarding balance dynamics, and discuss how we would look for signatures of balanced dynamics in convective systems. I'll also discuss the mechanism by which a vorticity anomaly can modulate and strengthen a developing convective system, and address the question of whether the Madden-Julian Oscillation is a candidate for a convective disturbance under the influence of balanced dynamics. Finally, I discuss how these concepts can potentially be used to evaluate and diagnose global models that have varying degrees of skill in simulating tropical disturbances (and the MJO in particular).
Special Wednesday Date--Rescheduled from 18 May 2017 Due to Weather
Wednesday, 24 May 2017, 3:30 PM
Refreshments 3:15 PM
NCAR--Foothills Laboratory
3450 Mitchell Lane
Bldg. 2, Main Auditorium, Room 1022
Anders Sivle
Norwegian Meteorological Institute
Oslo, Norway
Different people in different occupations depend on weather forecasts to plan their work and recreational schedules. People with no expertise in meteorology frequently interpret weather forecasts and uncertainty information. These non-experts apply their prior knowledge and experiences in a variety of fields to synthesize different types of information to interpret forecasts. In this PhD study, situations of typical users were simulated when examining how different user groups interpret, integrate, and use information from an online weather report (www.Yr.no) in their everyday decision-making. First, qualitative interviews of twenty-one Norwegians (farmers, exterior painters, tour guides, teachers and students) were conducted. Second, sixteen students participated in an eye-tracking study.
The study found that nuances such as color and the number of drops were important in the interpretations of the weather symbols and forecast uncertainty, which were sometimes interpreted differently than intended by the forecast provider. Prior knowledge and the integration of information from different representations affected the participants’ interpretations. The decision-making process influenced the selections of representations in different situations; their selection was dependent on the importance of the envisaged activity and the weather conditions for the day. Additionally, in situations in which the participants had a lack of experiences, this lack provides a possible explanation for why part of the information was occasionally not understood and used.
Some implications of the findings for communication and future research will be discussed in the presentation. For example, it appears that some users should be supported to facilitate the interpretation and use of information in situations where they lack experiences. One possibility to support persons that lack experiences and have low situation awareness might be to provide consequences and impacts of forecast weather.
Thursday, 8 June 2017, 3:30 PM
Refreshments 3:15 PM
NCAR-Foothills Laboratory
3450 Mitchell Lane
(Location Change) Bldg. 2, Room 1001
MMM Organizational Chart
MMM Staff & Visitors Directory
MMM Photo Gallery
MMM Laboratory Office
Phone: 303.497.8917
