Annual Experiments and Other Activities
In collaboration with the atmospheric and hydrologic forecasting and research communities, the HMT at WPC conducts an annual Flash Flood and Intense Rainfall Experiment for four weeks during the warm-season summer months. This experiment focuses on improving short term quantitative precipitation (QPF) and flash flood forecasts through the use of high resolution models and ensembles and rapidly updating hydrologic information.
In an effort to support improvements in winter weather forecasts, the HMT at WPC began hosting an annual winter weather experiment in 2011. These experiments bring together members of the operational forecasting, research, and academic communities to address winter weather forecast challenges. Previous experiments have focused on exploring the utility of high-resolution models, model implicit snowfall forecasting techniques to improve both deterministic and probabilistic snow and ice forecasts in the Day 1-7 time frame.
The NGGPS Day 8-10 Project is designed to establish a baseline skill at days 8-10 for existing operational datasets, assess the skill and utilty of experimental guidance to make daily forecasts in the day 8-10 period, evaluate new and experimental versions of NCEP guidance, and develop new day 8-10 forecast products which highlight high-impact hazardous events. The Day 8-10 Experiment is conducted remotely throughout the year with partners from the academic research community, the Climate Prediction Center (CPC), and the Environmental Modeling Center (EMC).
In collaboration with HMT at PSD and the Developmental Testbed Center (DTC), HMT at WPC is exploring the use of object-oriented verification. The goal of this type of verification is to better account for spatial discontinuities between the forecast and observed precipitation. We are using the Method for Object-Based Diagnostic Evaluation (MODE) tool for the verification of model and WPC forecaster QPF. The MODE tool is part of the Model Evaluation Tools (MET) verification package developed by DTC.
The 7th annual Flash Flood and Intense Rainfall (FFaIR) Experiment was conducted at the National Center for Weather and Climate Prediction (NCWCP) from June 17 through July 19, 2019 by the Hydrometeorology Testbed at the Weather Prediction Center (HMT-WPC). The HMT partnered with National Weather Service meteorologists and hydrologists and the development and research communities to evaluate the usefulness of various experimental model and ensemble products. In addition to the model data, various hydrological and satellite tools were used to produce short range (6-36 hour) probabilistic flash flood forecasts. The design of the experiment centered around mimicking operations at WPC and tasked participants each day to produce four experimental probabilistic forecasts utilizing the experimental tools provided by the HMT team. A secondary component of the experiment included analysis of the performance of convection allowing models (CAMs) that use the finite volume cubed-sphere dynamic core, referred to as the FV3 core. Experiment Summary
The 9th annual Winter Weather Experiment was conducted both remotely and on site at the National Center for Weather and Climate Prediction (NCWCP) from November 13, 2018 through March 15, 2019. The experiment focused on evaluating precipitation-type algorithms applied to both the FV3-GFS and 12-km NAM models during diverse weather events exhibiting precipitation-type transition zones and their use as inputs into a manually-generated experimental forecast blend. Ensemble predictability of winter weather events was also evaluated using an ensemble clustering tool. CAMs snowfall output from both deterministic and ensemble guidance for the Day 1 period were assessed, focusing on Great Lakes lake-effect snowfall and other mesoscale events. Experiment Summary
The sixth annual Flash Flood and Intense Rainfall Experiment (FFaIR) was conducted at the National Center for Weather and Climate Prediction (NCWCP) June 18-July 20, 2018. The 2018 FFaIR Experiment focused on the use of high resolution guidance to synthesize atmospheric and hydrological guidance in an end to end forecast process to produce probabilistic flash flood forecasts in the short range (6-24 hours). To simulate the flow of information that occurs from a national center (i.e. WPC) to the local forecast offices, this years experiment engaged the HMT-Hydro participants and the Science and Operations Officer-Development and Operations Hydrologist (SOO-DOH) community through screen sharing, video, and teleconference to discuss the experimental guidance with the goal of producing a collaborative 6-hour probabilistic flash flood forecast. There was significant emphasis on testing the high-resolution deterministic models and ensembles over the short term, Day 1, period and fusing the atmospheric data with new, experimental output from the National Water Model. Experiment Summary
The 8th Annual Winter Weather Experiment was executed both remotely and on site at the Weather Prediction Center in College Park, MD by the HMT-WPC from November 14, 2017 through March 9, 2018. The experiment objectives focused on the exploration and testing of multiple microphysical and probabilistic precipitation type methodologies to determine which methods best enhanced the forecast process. Tools and algorithms from NWP guidance, the WPC Winter Weather Desk, the National Blend of Models, and Fuzzy Clustering were evaluated. Experiment Summary
The fifth annual Flash Flood and Intense Rainfall Experiment (FFaIR) was conducted at the NOAA Center for Weather and Climate Prediction (NCWCP) June 19-July 21, 2017. The 2017 FFaIR Experiment focused on the use of high resolution atmospheric and hydrologic guidance to improve flood prediction both in the short range (6-12 hours) and at longer time ranges (48-72 hours). Virtual engagment with the NWS SOO community took place each day to enhance flood situational awareness and collaborate on potential experimental flash flood watches. There was significant emphasis on testing the improvement of the Days 2 and 3 probabilistic Excessive Rainfall Outlook (ERO) with higher resolution data and the fusion with the National Water Model output. Experiment Summary
The 7th Annual Winter Weather Experiment was hosted by HMT-WPC from January 17 through February 17, 2017. The experiment brought together forecasters, researchers, and model developers from across the weather enterprise to test new data sets, tools, and forecast methodologies. Building on last year's experiment, further testing of hourly snowfall rate data was conducted to assess the diagnosis and predictibility of meso-scale heavy snowfall banding. Several new PWPF methodologies were tested over the CONUS in the Day 2 period, and the experiment was rounded off by exploring the issuance of both leagacy and experimental winter weather watches from the national center. Forecasts and new tools were presented each session via a daily webinar. Experiment Summary
The fourth annual Flash Flood and Intense Rainfall Experiment (FFaIR) was conducted at the NOAA Center for Weather and Climate Prediction (NCWCP) June 20-July 22, 2016. The focus of FFaIR 2016 was to test and apply the use of flash flood tools from a national perspective down to a local Weather Forecast Office (WFO) level. Once again cross test-bed collaboration between WPC-HMT and HMT-Hydro focused on the goal of improving the forecasting of flash flooding. The use of OWP's National Water Model (NWM) coupled with atmospheric CAMs and ensembles were explored to test the issuance of Day 1 and 2 probabilistic flash flood forecasts. Experiment Summary | Addendum
HMT-WPC hosted the 6th Annual Winter Weather Experiment from January 28 through February 19, 2016. The first two days of the experiment were lost due to impact of the recent blizzard on the HMT-WPC team living in the local area. The experiment brought together forecasters, researchers, and model developers from across the weather enterprise to test new data sets and tools, with a focus on the utility of hourly snowfall rate forecasts. The utility of the WPC watch/warning collaboration tool was also explored each day. For the second year, a daily webinar featuring the experiment forecast for the day and the tools being tested was presented. For the first time, participants made a retrospective forecast using an archived event from earlier in the winter. Experiment Summary | Addendum | R2O Progress Table
The third annual Flash Flood and Intense Rainfall Experiment (FFaIR) was conducted at the NOAA Center for Weather and Climate Prediction (NCWCP) July 6-24, 2015. Cross test-bed collaboration between WPC-HMT and HWT-Hydro focused on the goal of improving the forecasting of flash floods. The utility of several high resolution models and ensembles systems was explored, as well as the Advective-Statistical Forecasts of Rainfall (ADSTAT) and forecast recurrence interval tools. Experiment Summary
From January 12 - February 13, HMT-WPC hosted it's annual Winter Weather Experiment. Now in it's fifth year, the experiment brought together forecasters, researchers, and model developers from across the weather enterprise to conduct exercises aimed at continuing improvement of winter weather forecasts. This year's focus of the experiment was the continued exploration of micro-physics based snowfall forecasting methods, along with expanding medium range probabilistic winter weather forecasts to include forecasts of heavy snow and freezing rain from Days 4-7. Additionally, this year's experiment featured a daily webinar, where participants shared their forecast thoughts and discussed experimental model data with remote participants. Experiment Summary
The second annual Flash Flood and Intense Rainfall Experiment (FFaIR) was conducted July 7-25, 2014. Building off of the success of last year's experiment, participants from across the weather enterprise explored the utility of neighborhood probability techniques and other convection-allowing model guidance to improve short-term flash flood forecasts. In addition to issuing experimental forecasts, participants in FFaIR provided a daily forecast briefing to support the HWT-Hydro experimental flash flood warning activities. FFaIR also featured the debut of a new flash flood verification database. The development of this database represents a significant step forward in our verification capabilities by combining flsh flood information from multiple sources. Experiment Summary
The fourth annual HMT-WPC Winter Weather Experiment was conducted January 21 - February 21, 2014. Experiment participants included forecasters, researchers, and model developers from a variety of sectors of the weather enterprise, and for the first time remote participation by National Weather Service forecasters from selected offices. The focus of the experment was the continued exploration of micro-physics based snowfall forecasting methods, along with a first time look at the coupling of NAM micro-physics with the Noah Land Surface Model to generate snowfall accumulation forecasts. Another first in this year's experiment was the exploration of medium range probablisitc winter weather forecasts. Participants evaluated several medium range guidance data sets, and generated Day 4-7 forecasts each day. Experiment Summary
In collaboration with the National Severe Storms Laboratory (NSSL) and the Earth System Research Laboratory (ESRL), HMT-WPC conducted the inaugural Flash Flood and Intense Rainfall Experiment (FFaIR) July 8–26, 2013. During the experiment, forecasters, researchers, and model developers worked together as a collaborative forecast team to explore the challenges associated with issuing short-term QPF and flash flood forecasts. The experiment featured a variety of deterministic and probabilistic high resolution model data and identified several new forecast tools that will be used on WPC's new MetWatch Desk. Experiment Summary
The third annual HMT-HPC Winter Weather Experiment was conducted January 15 - February 15, 2013. The experiment brought together forecasters, researchers, and model developers from across the weather enterprise to explore the use of ensemble systems to help quantify and communicate uncertainty in winter weather forecasts. This year's experiment also highlighted a new snowfall accumulation technique that uses model microphysics information to refine the snowfall forecasts. The experiment highlighted the importance of ensemble guidance for winter weather forecasts and identified several potential ways to improve and expand WPC's current winter weather product suite. Experiment Summary
The first Atmospheric River Retrospective Forecasting Experiment (ARRFEX) was a collaboration between the Earth System Research Laboratory (ESRL) and the Hydrometeorological Testbed (HMT)—Hydrometeorological Prediction Center (HPC). The experiment was held September 17-28, 2012 at the new National Center for Weather and Climate Prediction (NCWCP). The experiment brought together forecasters, modelers and researchers from River Forecast Centers (RFC), Weather Forecast Offices (WFO), the Environmental Modeling Center (EMC), ESRL and the University of Utah to identify potential techniques to improve forecasts of atmospheric river (AR) induced extreme precipitation events along the U.S. West Coast. The experiment featured retrospective analysis of 8 pre-selected AR events that resulted in heavy precipitation along the U.S. West Coast during the 2009-2012 cool seasons. Participants were instructed to create a variety of forecast products, in pseudo-real time, for each event using archived operational and experimental numerical model guidance and datasets. Experiment Summary
The second annual HMT-HPC Winter Weather Experiment was conducted January 9 – February 10, 2012, and brought together forecasters, researchers and model developers to explore the challenges of probabilistic winter weather forecasting. The experiment focused on the use of ensemble systems to help quantify and communicate uncertainty in winter weather forecasts. The experiment revealed that while probabilistic winter weather forecasting remains a challenge, higher resolution ensembles can provide critical forecast details needed to isolate specific parameters for better depiction of forecast uncertainty. Experiment Summary
Continuing the partnership started in 2010, HMT-HPC staff and HPC forecasters once again traveled to Norman, OK to facilitate the QPF component of the 2011 HWT Spring Experiment. During the experiment, participants evaluated the utility of high resolution deterministic and ensemble guidance for precipitation forecasting. High resolution ensembles were again found to provide skillful forecast guidance, highlighting the potential for this type of guidance to add further skill to warm season QPF in the future. Experiment Summary
The inuagural HMT-HPC Winter Weather Experiment was held at HPC during January-February 2011. HPC forecasters as well as visiting scientists from NOAA-HMT, other NCEP centers and NWS field officies participated in the experiment. The primary focus of the experiment was to explore the role of high-resolution model data in improving forecasts of precipitation type, snow and ice accumulation, and meso-scale snow bands. The experiment revealed that the high resolution models were often too slow with the progression of the surface low, which impacted their precipitation and thermal fields. Experiment Summary
HMT-HPC partnered with the Hazardous Weather Testbed to add a QPF component to their 2010 Spring Experiment. During the experiment, participants explored the use of high resolution convection-allowing deterministic and ensemble guidance for potential improvement of warm season QPF. High resolution models were found to have skill in forecasting warm season QPF, which portends well for future forecast improvements. Forecast guidance from the 4km, 26 member Storm Scale Ensemble Forecast system (SSEF) provided by University of Oklahoma's Center for Analysis and Prediction of Storms (CAPS) was particularly impressive. The QPF Component of the 2010 HWT Spring Experiment was presented at the AMS Annual Meeting in Seattle, WA. In addition to leading the QPF component of the HWT Spring Experiment, HMT–HPC hosted a two week parallel experiment for HPC forecasters. Related Publication
HPC collaborated with researchers from Texas A&M University as part of a COMET Partners project that focused on determining whether operational numerical weather prediction (NWP) models have a consistent displacement bias in the predicted location of elevated warm season convective systems. Using the Method for Object-Based Diagnostic Evaluation (MODE) tool, Texas A&M researchers objectively confirmed forecasters' subjective impression that there is a north bias in the location of elevated convection systems in both the NAM and the GFS.
WPC is experimenting with probabilisitc QPF products. The products consist of both the probability of exceeding a given precipitation threshold as well as precipitation amounts by percentile. The products are generated automatically using the WPC deterministic QPF and uncertainty information from an ensemble. A binormal probability distribution (density) function (PDF), which allows skewness, is constructed such that the mode is the WPC QPF and the variance is that of the ensemble. The skewness is based on the position of the WPC QPF in the ensemble distribution. This approach to estimating the three parameters for the binormal PDF is a variation on the method of moments.
HMT at WPC is investigating the diagnosis and identification of predecessor rainfall events (PREs). A comparison of model performance between the NAM and the GFS as well as an evaluation of WPC QPF is being looked at to help improve WPC QPF forecasts. The application of a meridional moisture flux tool is being investigated to identify transport of deep tropical moisture into the mid-latitude heavy rainfall system.
HMT at WPC collaborated with EMC cloud physicist Brad Ferrier to develop an algorithm to estimate hourly snowfall accumulation. The algorithm utilizes a parameter in the Ferrier microphysics scheme called the "rime factor" to estimate one-hour instantaneous snow/ice accumulation from NCEP's NAM 12km model. This data was made available to WPC forecasters in NAWIPS. (Fall 2009)
Training was provided to WPC Forecasters during November 2009.
In November 2012, an updated algorithm was delivered to NAWIPS. The update used the rime-facter data from the NAM12 to modify the Roebber Snow to Liquid Ratio (SLR) computed for the NAM. This modified SLR was then multiplied by the percent of frozen precipitation output from the NAM12, and the NAM12 post processed precipitation to arrive at a snow accumulation forecast. This data set was tested in the 2013 Winter Weather Experiment and implemented into WPC winter weather operations in February 2013.
For forecasters preparing QPFs. This training material is available online and contains a wealth of information on warm season precipitation forecasting. (Summer 2008)
HMT at WPC evaluated the Whitaker-Hamill reforecast technique and migrated it to NCEP computers. This data is now available for WPC forecasters in NAWIPS to aid in precipitation forecasting. (Winter 2006-2007)
The use of normalized anomalies has been successfully migrated into WPC operations. Model forecasts of normalized anomalies for both mass fields and mositure diagnostics (precipitable water and moisture flux) are available both in NAWIPS and online. (Winter 2005-2006)