Browse

You are looking at 91 - 100 of 2,838 items for :

  • Weather and Forecasting x
  • Refine by Access: All Content x
Clear All
Florian Dupuy, Olivier Mestre, Mathieu Serrurier, Valentin Kivachuk Burdá, Michaël Zamo, Naty Citlali Cabrera-Gutiérrez, Mohamed Chafik Bakkay, Jean-Christophe Jouhaud, Maud-Alix Mader, and Guillaume Oller

Abstract

Cloud cover provides crucial information for many applications such as planning land observation missions from space. It remains, however, a challenging variable to forecast, and numerical weather prediction (NWP) models suffer from significant biases, hence, justifying the use of statistical postprocessing techniques. In this study, ARPEGE (Météo-France global NWP) cloud cover is postprocessed using a convolutional neural network (CNN). CNN is the most popular machine learning tool to deal with images. In our case, CNN allows the integration of spatial information contained in NWP outputs. We use a gridded cloud cover product derived from satellite observations over Europe as ground truth, and predictors are spatial fields of various variables produced by ARPEGE at the corresponding lead time. We show that a simple U-Net architecture (a particular type of CNN) produces significant improvements over Europe. Moreover, the U-Net outclasses more traditional machine learning methods used operationally such as a random forest and a logistic quantile regression. When using a large number of predictors, a first step toward interpretation is to produce a ranking of predictors by importance. Traditional methods of ranking (permutation importance, sequential selection, etc.) need important computational resources. We introduced a weighting predictor layer prior to the traditional U-Net architecture in order to produce such a ranking. The small number of additional weights to train (the same as the number of predictors) does not impact the computational time, representing a huge advantage compared to traditional methods.

Open access
Ying Wang and Zhaoxia Pu

Abstract

The benefits of assimilating Next Generation Weather Radar (NEXRAD) radial velocity data for convective systems have been demonstrated in previous studies. However, impacts of assimilation of such high spatial and temporal resolution observations on hurricane forecasts have not been demonstrated with the National Centers for Environmental Prediction (NCEP) Hurricane Weather and Research Forecasting (HWRF) system. This study investigates impacts of NEXRAD radial velocity data on forecasts of the evolution of landfalling hurricanes with different configurations of data assimilation. The sensitivity of data assimilation results to influencing parameters within the data assimilation system, such as the maximum range of the radar data, superobservations, horizontal and vertical localization correlation length scale, and weight of background error covariances, are examined. Two hurricane cases, Florence and Michael, which occurred in the summer of 2018 are chosen to conduct a series of experiments. Results show that hurricane intensity, asymmetric structure of inland wind and precipitation, and quantitative precipitation forecasting are improved. Suggestions for implementation of operational configurations are provided.

Open access
Hailiang Du

Abstract

The evaluation of probabilistic forecasts plays a central role both in the interpretation and in the use of forecast systems and their development. Probabilistic scores (scoring rules) provide statistical measures to assess the quality of probabilistic forecasts. Often, many probabilistic forecast systems are available while evaluations of their performance are not standardized, with different scoring rules being used to measure different aspects of forecast performance. Even when the discussion is restricted to strictly proper scoring rules, there remains considerable variability between them; indeed strictly proper scoring rules need not rank competing forecast systems in the same order when none of these systems are perfect. The locality property is explored to further distinguish scoring rules. The nonlocal strictly proper scoring rules considered are shown to have a property that can produce “unfortunate” evaluations, particularly the fact that the continuous rank probability score prefers the outcome close to the median of the forecast distribution regardless of the probability mass assigned to the value at/near the median raises concern to its use. The only local strictly proper scoring rule, the logarithmic score, has direct interpretations in terms of probabilities and bits of information. The nonlocal strictly proper scoring rules, on the other hand, lack meaningful direct interpretation for decision support. The logarithmic score is also shown to be invariant under smooth transformation of the forecast variable, while the nonlocal strictly proper scoring rules considered may, however, change their preferences due to the transformation. It is therefore suggested that the logarithmic score always be included in the evaluation of probabilistic forecasts.

Restricted access
Carlo Cafaro, Beth J. Woodhams, Thorwald H. M. Stein, Cathryn E. Birch, Stuart Webster, Caroline L. Bain, Andrew Hartley, Samantha Clarke, Samantha Ferrett, and Peter Hill

Abstract

Convection-permitting ensemble prediction systems (CP-ENS) have been implemented in the midlatitudes for weather forecasting time scales over the past decade, enabled by the increase in computational resources. Recently, efforts are being made to study the benefits of CP-ENS for tropical regions. This study examines CP-ENS forecasts produced by the Met Office over tropical East Africa, for 24 cases in the period April–May 2019. The CP-ENS, an ensemble with parameterized convection (Glob-ENS), and their deterministic counterparts are evaluated against rainfall estimates derived from satellite observations (GPM-IMERG). The CP configurations have the best representation of the diurnal cycle, although heavy rainfall amounts are overestimated compared to observations. Pairwise comparisons between the different configurations reveal that the CP-ENS is generally the most skillful forecast for both 3- and 24-h accumulations of heavy rainfall (97th percentile), followed by the CP deterministic forecast. More precisely, probabilistic forecasts of heavy rainfall, verified using a neighborhood approach, show that the CP-ENS is skillful at scales greater than 100 km, significantly better than the Glob-ENS, although not as good as found in the midlatitudes. Skill decreases with lead time and varies diurnally, especially for CP forecasts. The CP-ENS is underspread both in terms of forecasting the locations of heavy rainfall and in terms of domain-averaged rainfall. This study demonstrates potential benefits in using CP-ENS for operational forecasting of heavy rainfall over tropical Africa and gives specific suggestions for further research and development, including probabilistic forecast guidance.

Open access
Akila Sampath, Uma S. Bhatt, Peter A. Bieniek, Robert Ziel, Alison York, Heidi Strader, Sharon Alden, Richard Thoman, Brian Brettschneider, Eugene Petrescu, Peitao Peng, and Sarah Mitchell

Abstract

In this study, seasonal forecasts from the National Centers for Environmental Prediction (NCEP) Climate Forecast System, version 2 (CFSv2), are compared with station observations to assess their usefulness in producing accurate buildup index (BUI) forecasts for the fire season in Interior Alaska. These comparisons indicate that the CFSv2 June–July–August (JJA) climatology (1994–2017) produces negatively biased BUI forecasts because of negative temperature and positive precipitation biases. With quantile mapping (QM) correction, the temperature and precipitation forecasts better match the observations. The long-term JJA mean BUI improves from 12 to 42 when computed using the QM-corrected forecasts. Further postprocessing of the QM-corrected BUI forecasts using the quartile classification method shows anomalously high values for the 2004 fire season, which was the worst on record in terms of the area burned by wildfires. These results suggest that the QM-corrected CFSv2 forecasts can be used to predict extreme fire events. An assessment of the classified BUI ensemble members at the subseasonal scale shows that persistently occurring BUI forecasts exceeding 150 in the cumulative drought season can be used as an indicator that extreme fire events will occur during the upcoming season. This study demonstrates the ability of QM-corrected CFSv2 forecasts to predict the potential fire season in advance. This information could, therefore, assist fire managers in resource allocation and disaster response preparedness.

Restricted access
Nathan A. Wendt and Israel L. Jirak

Abstract

The Multi-Radar Multi-Sensor (MRMS) system generates an operational suite of derived products in the National Weather Service useful for real-time monitoring of severe convective weather. One such product generated by MRMS is the maximum estimated size of hail (MESH) that estimates hail size based on the radar reflectivity properties of a storm above the environmental 0°C level. The MRMS MESH product is commonly used across the National Weather Service (NWS), including the Storm Prediction Center (SPC), to diagnose the expected hail size in thunderstorms. Previous work has explored the relationship between the MRMS MESH product and severe hail (≥25.4 mm or 1 in.) reported at the ground. This work provides an hourly climatology of severe MRMS MESH across the contiguous United States from 2012 to 2019, including an analysis of how the MESH climatology differs from the severe hail reports climatology. Results suggest that the MESH can provide beneficial hail risk information in areas where population density is low. Evidence also shows that the MESH can provide potentially beneficial information about severe hail occurrence during the night in locations that are climatologically favored for upscale convective growth and elevated convection. These findings have important implications for the use of MESH as a verification dataset for SPC probabilistic hail forecasts as well as severe weather watch decisions in areas of higher hail risk but low population density.

Restricted access
Benjamin A. Schenkel, Michael Coniglio, and Roger Edwards

Abstract

This work investigates how the relationship between tropical cyclone (TC) tornadoes and ambient (i.e., synoptic-scale) deep-tropospheric (i.e., 850–200-hPa) vertical wind shear (VWS) varies between coastal and inland environments. Observed U.S. TC tornado track data are used to study tornado frequency and location, while dropsonde and radiosonde data are used to analyze convective-scale environments. To study the variability in the TC tornado–VWS relationship, these data are categorized by both 1) their distance from the coast and 2) reanalysis-derived VWS magnitude. The analysis shows that TCs produce coastal tornadoes regardless of VWS magnitude primarily in their downshear sector, with tornadoes most frequently occurring in strongly sheared cases. Inland tornadoes, including the most damaging cases, primarily occur in strongly sheared TCs within the outer radii of the downshear-right quadrant. Consistent with these patterns, dropsondes and coastal radiosondes show that the downshear-right quadrant of strongly sheared TCs has the most favorable combination of enhanced lower-tropospheric near-surface speed shear and veering, and reduced lower-tropospheric thermodynamic stability for tornadic supercells. Despite the weaker intensity farther inland, these kinematic conditions are even more favorable in inland environments within the downshear-right quadrant of strongly sheared TCs, due to the strengthened veering of the ambient winds and the lack of changes in the TC outer tangential wind field strength. The constructive superposition of the ambient and TC winds may be particularly important to inland tornado occurrence. Together, these results will allow forecasters to anticipate how the frequency and location of tornadoes and, more broadly, convection may change as TCs move inland.

Restricted access
Christopher J. Schultz, Roger E. Allen, Kelley M. Murphy, Benjamin S. Herzog, Stephanie A. Weiss, and Jacquelyn S. Ringhausen

Abstract

Infrequent lightning flashes occurring outside of surface precipitation pose challenges to Impact-Based Decision Support Services (IDSS) for outdoor activities. This paper examines the remote sensing observations from an event on 20 August 2019 where multiple cloud-to-ground flashes occurred over 10 km outside surface precipitation (lowest radar tilt reflectivity < 10 dBZ and no evidence of surface precipitation) in a trailing stratiform region of a mesoscale convective system. The goal is to demonstrate the fusion of radar with multiple lightning observations and a lightning risk model to demonstrate how reflectivity and differential reflectivity combined provided the best indicator for the potential of lightning where all of the other lightning safety methods failed. A total of 13 lightning flashes were observed by the Geostationary Lightning Mapper (GLM) within the trailing stratiform region between 2100 and 2300 UTC. The average size of the 13 lightning flashes was 3184 km2, with an average total optical energy of 7734 fJ. A total of 75 NLDN flash locations were coincident with the 13 GLM flashes, resulting in an average of 5.8 NLDN flashes [in-cloud (IC) and cloud-to-ground (CG)] per GLM flash. In total, five of the GLM flashes contained at least one positive cloud-to-ground flash (+CG) flash identified by the NLDN, with peak amplitudes ranging between 66 and 136 kA. All eight CG flashes identified by the NLDN were located more than 10 km outside surface precipitation. The only indication of the potential of these infrequently large flashes was the presence of depolarization streaks in differential reflectivity (Z DR) and enhanced reflectivity near the melting layer.

Restricted access
Jonathan Poterjoy, Ghassan J. Alaka Jr., and Henry R. Winterbottom

Abstract

Limited-area numerical weather prediction models currently run operationally in the United States and follow a “partially cycled” schedule, where sequential data assimilation is periodically interrupted by replacing model states with solutions interpolated from a global model. While this strategy helps overcome several practical challenges associated with real-time regional forecasting, it is no substitute for a robust sequential data assimilation approach for research-to-operations purposes. Partial cycling can mask systematic errors in weather models, data assimilation systems, and data preprocessing techniques, since it introduces information from a different prediction system. It also adds extra heuristics to the model initialization steps outside the general Bayesian filtering framework from which data assimilation methods are derived. This study uses a research-oriented modeling system, which is self-contained in the operational Hurricane Weather Research and Forecasting (HWRF) Model package, to illustrate why next-generation modeling systems should prioritize sequential data assimilation at early stages of development. This framework permits the rigorous examination of all model system components—in a manner that has never been done for the HWRF Model. Examples presented in this manuscript show how sequential data assimilation capabilities can accelerate model advancements and increase academic involvement in operational forecasting systems at a time when the United States is developing a new hurricane forecasting system.

Restricted access
Gregory J. Stumpf and Alan E. Gerard

Abstract

Threats-in-Motion (TIM) is a warning generation approach that would enable the NWS to advance severe thunderstorm and tornado warnings from the current static polygon system to continuously updating polygons that move forward with a storm. This concept is proposed as a first stage for implementation of the Forecasting a Continuum of Environmental Threats (FACETs) paradigm, which eventually aims to deliver rapidly updating probabilistic hazard information alongside NWS warnings, watches, and other products. With TIM, a warning polygon is attached to the threat and moves forward along with it. This provides more uniform, or equitable, lead time for all locations downstream of the event. When forecaster workload is high, storms remain continually tracked and warned. TIM mitigates gaps in warning coverage and improves the handling of storm motion changes. In addition, warnings are automatically cleared from locations where the threat has passed. This all results in greater average lead times and lower average departure times than current NWS warnings, with little to no impact to average false alarm time. This is particularly noteworthy for storms expected to live longer than the average warning duration (30 or 45 min) such as long-tracked supercells that are more prevalent during significant tornado outbreaks.

Open access