A Pilot Study Examining Model Derived Precipitation Efficiencyfor Use in Precipitation Forecasting in the Eastern United States

 

James Noel

NOAA/NWS Forecast Office, Atlanta, Georgia

 

Jeffrey C. Dobur

NOAA/NWS Forecast Office, Atlanta, Georgia

 

 

 

ABSTRACT

 

   In 1996, the Ohio River Forecast Center(OHRFC) implemented model-derived Precipitation Efficiency to assist inevaluating the expected spatial and temporal distribution ofprecipitation.   Precipitation Efficiency(PE) is derived from any atmospheric numerical weather prediction model whereprecipitable water for the entire atmospheric column and mean relative humidityfor the 1000 to 700 hPa layer are computed. The goal of this paper is to describe a technique by which precipitationforecasting skills can be improved using PE.

                The PEmodel-derived parameter has proven to be a useful tool in refining the probability,timing, duration, coverage, and intensity of precipitation.  PE was shown to provide value-addedinformation to assist the hydrometeorologist in preparing precipitationforecasts.  Results using the NationalWeather Service’s (NWS) Eta model show that as the value of PE increases, the percentof precipitation occurrences also increases.  In addition, results indicate that the onset of precipitation is tied tocritical PE values and temperatures.   

 

 

 

 

 

 

1.  Introduction

 

QuantitativePrecipitation Forecasts (QPF) have been an integral part of the river and floodforecast program in the National Weather Service (NWS) Eastern Region since1977 (Opitz et al. 1995) and NWS-wide since the 1990s (Fenbers 1995).  Forecasts of precipitation amounts and onsetare critical to the achievement of the greatest possible hydrologic forecastaccuracy and longest possible lead-times (Georgakakus and Hudlow 1984).   As part of the National Weather Service(NWS) modernization program, a Hydrometeorological Analysis and Support (HAS)function was created at the NWS River Forecast Centers (RFC) to maintain theQPF process.  The HAS function utilizesthe 6-hour national QPF guidance from the Hydrometeorological Prediction Center(HPC) in addition to examining an array of meteorological model and mesoscaleparameters in formulating the 12-24-hour HAS QPF.   The HAS QPF is completed twice daily around0000 UTC and 1200 UTC and is incorporated into the NWS River Forecast System(NWSRFS) to produce river forecasts out to three to five days.  In addition toaddressing the spatial and temporal challenges of precipitation forecasting atRFCs, there is a  continuing need to improvethe Probability of Precipitation (POP) forecasts at NWS Weather Forecastoffices (WFOs).  Improved methods forprecipitation forecasting could benefit both NWS RFCs and WFOs.  One such method is presented here.

 

2.    Background

Precipitable water (PW) andmean relative humidity (RH) have been derived using real-time satelliteestimates at the National Oceanic and Atmospheric Administration (NOAA)National Environment, Satellite, Data and Information Service since the early1980s.  These parameters are then used toestimate how efficient the precipitation process is and adjustments to rainfallestimates can be subsequently made (NESDIS) (Scofield 1987; Vicente andScofield 1998).  In 1996, the OHRFC appliedthe NESDIS PW/RH method to model-derived precipitation forecasts.  In order to apply a real-time technique tomodel-based forecasts, there was a need to use model-derived weather parametersto approximate precipitation efficiency (PE).  Precipitation Efficiency is defined as theratio of the total rainfall to the total condensation (Weisman and Klemp 1982and Ferrier et al. 1996).  While theformer can be derived from standard numerical models, the latter is notavailable.  Another approximation isneeded.

Several factorsaffect PE, including saturation ratio, production rate of condensate, residencetime of droplets in clouds, dry air entrainment, vertical wind shear, andprecipitable water (Doswell et. al. 1996).  A requisite for high rainfall intensity is a large production rate ofcondensate.  The rate at which thecondensate is produced in a column of cloudy air is directly proportional toair density, updraft speed, cloud thickness, and the vertical gradient of thesaturation mixing ratio. The density and vertical gradient of the saturationmixing ratio terms act to produce larger condensate rates in the lower part ofthe cloudy column.   In general, thegreatest rates of condensate production are found in the lower half of thecloudy column.  The residence time ofdroplets in clouds also plays a critical role in increasing PE.  With increased vertical motion and increaseddepth of clouds, cloud droplets are allowed longer residence time in the cloudto grow large enough to produce rain droplets. Vertical wind shear plays a critical role since the shear often producesdry air entrainment, reducing PE. Finally, a high amount of precipitable water usually increases PE.  Typically, precipitable water values rangefrom 1.50” and greater during the warm season (Junker 1997) to around 0.80” ormore in the cool season. 

 For the operational hydrometeorologist, asimple relationship related to precipitation efficiency is defined as PE = PW xRH; where RH is the average lower tropospheric relative humidity and PW is theprecipitable water through the entire column (Scofield 1987).  PW and RH are easily obtainable from numericalweather prediction models, although  theusefulness of gridded data is limited by the model from which it is derived (Scofieldand Kusselson 1996).  The PW/RH relationshipindicates a potential efficiency of the environment for producing precipitationat specific times in the future.  Thus,this model-derived PW/RH parameter is referred to as Precipitation Efficiency (PE)for the operational forecast process at NWS WFOs and RFCs.  It must be emphasized that this model-derivedPE is only an approximation of PE, using only PW and RH, not actual PE,discussed earlier. 

 

3. Data and methodology

 

This sectiondiscusses ways PE can be implemented into the precipitation forecasting process,and a description of data sources and analysis techniques used.

 

PE is calculated asfollows:

 

PE = PW * (1000-700 MRH)

 

where PE = Precipitation Efficiency,PW = Precipitable Water through the entire depth of the atmosphere, in inches,and 1000-700 MRH is the mean relative humidity expressed as a decimalvalue.  The 1000-700 hPa layer was chosensince the deep moisture is mainly contained in the lowest 3-4 km of theatmosphere (Junker 1997). 

PE can be displayed as anadded volume browser customization within the Advanced Weather InteractiveProcessing System (AWIPS) D2D meteorological display software (Biere1998).  Readily-available software suchas General Meteorological Data Assimilation, Analysis and Display SoftwarePackage (GEMPAK) (desJardins 1985), NWS National Centers Translator (Ntrans),and GEMPAK Analysis and Rendering Program (GARP) are also capable ofintegrating PE into their list of precipitation forecasting parameters.   This allows for widespread use of the PEparameter in all sectors (government, university, and private).

We examined twenty-sevencases from March 1997 through June 1998 (Table 1) in which precipitationoccurred within the OHRFC hydrologic service area.  An additional four cases occurring from May2001 through September 2001 were examined. 

In thecases for March 1997 through June 1998, PE values were taken from the NWS NCEP50 layer, 29-km Eta numerical weather model using GARP.   PE values were determined for each six-hour intervalof the 48-hour model forecast for six different cities in the Ohio Valley region (Table 2). Six-hourintervals were chosen due to the limits of model output intervals and resolutionat the time of data collection. In addition, six-hour intervals allow you tocapture model temporal and spatial uncertainty. Furthermore, occurrence of precipitation forecasts are usually made insix-hour intervals or greater. This dataset provided a total of 1296 forecasttimes and locations against which observed precipitation could be compared. Theseforecast values were compared to the monthly Local Climatological Data (LCD)hourly rainfall amounts at each location.  In addition, a warm season case from June 29th, 2001 and a transition season casefrom March 9th, 2002are shown.  Using the NWS AWIPS D2D meteorologicalanalysis software, a comparison was made between PE and the Ohio Valley regional 0.5°reflectivity radar mosaic to show the utility of PE.

            The PE values (inches) derived fromthe Eta model were compared against the percentage of observed precipitation occurrences(PPO).  The PPO was calculated by dividingthe number of occurrences by the total number of 6-hourly intervals for eachlocation.   An occurrence is defined aswhen 0.01” precipitation or greater was recorded at a particular location duringany hour of the 6-hour interval. The 6-hour intervals were grouped into threecategories to account for seasonal moisture influences driven by temperatureand amount of available moisture in the atmospheric column.   To do this, a mean temperature wascalculated for all the 6-hour intervals from May 1997 to June 1998. The first category,called the mean transition season category, was defined as those 6-hourintervals with a mean surface temperature within one standard deviation of the overallmean surface temperature (54°F). The second category, called the cool season category,was defined as those 6-hour intervals with a mean surface temperature more thanone standard deviation cooler than the overall mean.  The final category, called the warm season category,was defined as those 6-hourly intervals with a mean surface temperature morethan one standard deviation warmer than the overall mean temperature.  A linear regression line was computed for allgroups (Figure 1).

4. Results

 

In Figure 1, results showthe plot of the three categories linear regression lines. The PE valuesassociated with the 80 PPO for the cool, transition, and warm season categorieswere 0.75”, 1.15”, and 1.90”, respectively.  The PE value associated with the 50 PPO forthe cool, transition, and warm season categories were 0.50”, 0.75”, and 1.30”,respectively.  The PE value associatedwith the 20 PPO for the cool, transition, and warm season categories were 0.25”,0.30”, and 0.65”, respectively.  Thesedifferences between cool, transition, and warm season categories can be attributedto seasonal variations in moisture and the random nature of scattered afternoonconvection, especially during the warm season.  Correlation coefficients for the cool, transition, and warm season of 0.93,0.92, 0.90 respectively, provide confidence in the utility of this parameter. Basedon these results, OHRFC and WFO ATL have developed monthly precipitationthresholds (Figure 2). 

            PE has also shown the capability to indicateprecipitation intensity.  High values ofPW and instability are often collocated and become antecedent conditions priorto the development of heavy rainfall and flash floods (Scofield et al. 1996,2000).  High values of PW can producehigh values of PE if 1000-700 MRH is high. Data from March 1997 through June 1998 show evidence that the proportionof heavier precipitation occurrences (greater than 0.25” in a 6-hour period) tototal occurrences is larger with higher PE values (Figure 3). 

            In addition to providing some levelof confidence in the PPO, PE has displayed the ability to detail the axis ofprecipitation development and movement. This is especially important when a precipitation forecast is made forinput into a hydrologic model.  Spatiallycentering the axis of precipitation is critical in projecting which locationson certain rivers will rise, recess, or remain steady.  During times of high flow, such a prognosisin determining the axis of precipitation can mean the difference between issuingand not issuing a flood forecast.

            Comparisons of PE to radarreflectivity during the summer of 2001 and spring of 2002 have shown theability of PE values to highlight the axis and areal coverage ofprecipitation.  Spring and summer caseswere chosen to show PE performance in both a synoptic case (spring) and a localforcing case (summer).  On June 29, 2001,scattered convection developed in an area from Indianapolis, Indiana eastwardto near Dayton, Ohio during the morning hours.  Evaluating the 0000 UTC June 29th Eta  PE forecast valid at 1200 UTC overlaid withthe 1130 UTC  June 29th OhioValley regional 0.5°reflectivity mosaic (Figure 4), it is evident that PE provided a fair solutionin portraying the areal coverage and axis of precipitation.  In examining the 1200 UTC Eta PE forecastvalid at 1800 UTC overlaid with the 1800 UTC June 29th Ohio Valley regional 0.5° reflectivity mosaic (Figure 5),  PE provided an indication ofthe shift in developing convection by the afternoon across the Cumberland rivervalley.   PE performance is illustratedin a third example dealing with a cold front pushing across the Ohioand Tennessee valleys on March 9th, 2002.   PE 1200 UTC Eta-model values for the 1800 UTCperiod highlighted the impending coverage and axis of precipitation when comparedto the 0.5°reflectivity near the same time period (Figure 6). 

Finally, Figure 7 shows howPE can be used to better forecast precipitation than the individual componentsused to derive it. The four panel image displays Eta-model derived PW, Eta-model PE, Eta-model 1000-700 hPa mean RH, and anIR picture for 0600 UTC  March 30th 2002.  The main convection was occurring acrossnorthern Mississippi, Alabamaand Georgiainto eastern Tennessee.  PW focused on northern Mississippiand Alabama while 1000-700 hParelative humidity focused over eastern Tennessee.When combining PW/RH together, PE focused on the area where the strongconvection occurred.

 

 

5. Conclusions and Future Research

 

            PE is not a stand-alone indicatorfor precipitation, but it has been proven as a very useful tool in evaluating thespatial and temporal distribution of precipitation.  This parameter can assist in refiningprobability of precipitation forecasts. When applied alongside other traditional or useful parameters such as950-850 hPa low level jet convergence, 300-200 hPa upper level jet divergence,950-850 hPa theta-e advection, 850-500 hPa omega, and other indices, PE can bea more valuable tool than relying on its foundational components individually. 

            Additional case studies are neededto further examine the threshold criteria for heavy rainfall during differenttimes of the year and at different surface temperatures and dewpoints.  Further study will substantiate additionalvalue in using the PE parameter for other regions across the United States. 

 

 

Acknowledgments.  The authors would like to thank Rod Scofieldof NESDIS, Gary Beeley from WFO ATL, Wes Junker and Charlie Chappell from COMETtraining, and Dave Ondrejik from WFO CTP for ideas on improving the parameterand manuscript and to the reviewers for their suggestions.  A special thanks also goes to Mark Fenbersfrom OHRFC for the founding idea to implement PE into a forecast mode.

 

 

 

Authors

 

James Noel currently serves as the Senior ServiceHydrologist at the National Weather Service Forecast Office (WFO) in Peachtree City, Georgia.  Prior to this, he served as a Public andAviation Forecaster at the WFO Peachtree City, GA, a Hydrometeorologist at the Ohio River Forecast Centerin Wilmington, Ohio,and Developmental Meteorologist at the Techniques Development Lab at NationalWeather Service Headquarters in Silver Spring, Maryland.  His education includes a BS in Meteorologywith a minor in Math from Northern Illinois University (1992) and studies inHydrology/Civil Engineering from the Ohio State University.

 

Jeff Dobur currently serves as a Public and AviationForecaster at the National Weather Service Forecast Office (NWSFO) in Peachtree City, Georgia.  Prior to this, he has worked at the NWSFO andthe Ohio River Forecast Center in Wilmington, Ohio and with the State Climatologist in Ohio.  His educational background includes a BS inAtmospheric Science from the Ohio State University (1997) and additionalgraduate course work in Hydrology/Civil Engineering from the University of Cincinnati. 

 

 

 

 

 

 

REFERENCES

 

 

Biere, M.R., 1998: The WFO-Advanced Two-Dimensional DisplaySoftware Design. Preprints, Fourteenth

     InternationalConference on Interactive Information and Processing Systems (IIPS) forMeteorology,

     Oceanography, and Hydrology, 11 - 16 January1998, Phoenix, AZ, 376-379.

 

desJardins, M.L., and R. A. Petersen, 1985: GEMPAK: A meteorological system for research and

     education. Preprints, First InternationalConference on Interactive Information and Processing Systems

     for Meteorology, Oceanography, andHydrology, Los Angeles, Amer. Meteor.Soc., 313-319.

 

Doswell, C. A., H.E. Brooks and R.A. Maddox, 1996:  Flash Flood Forecasting: An Ingredients-Based         Methodology.  Weather and Forecasting, 11 560-581

 

Fenbers, M., 1995:  WINQPF. NOAA Eastern Region Computer Programs, NWS ERCP25MC, National

     Weather Service, NOAA, U.S. Dept. of Commerce.

 

Ferrier, B.S., J. Simpson,and Wei-Kuo Tao, 1996: Factors Responsible for Precipitation Efficiencies in

     Mid-latitude and tropical simulations.Mon. Wea. Rev., 124, 2100-2125.

 

Georgakakos K. and M. Hudlow, 1984:  Quantitative Precipitation ForecastTechniques for Use in

    Hydrologic Forecasting.  Bull.American Meteorological Society, 65,1186-1200.

 

Opitz, H.H., and Coauthors, 1995:  The Challenge of Forecasting Heavy Rain andFlooding throughout the Eastern Region of the National Weather Service.  Part II: Forecasting Techniques andApplications.  Weather and Forecasting,10 91-104.

 

ScofieldR.A., 1987: The NESDIS Operational Convective Precipitation EstimationTechnique. Mon. Wea.

     Rev., 115, 1773-1792.

 

Scofield, R.A. and S.Kusselson, 1996:  QuantitativePrecipitation Forecasting (QPF) – The End-To-End

     Forecasting Process Using Satellite dataand Numerical Weather Prediction Models. Proc. Fifteenth

     Conf. On Weather Analysis and Forecasting,Norfolk, V.A., Amer. Meteor. Soc., J142-J145.

 

Scofield R.A.., G.Vicenteand M. Hodges, 2000 :  The Use of WaterVapor for Detecting Environments

     that Lead to Convectively Produced HeavyPrecipitation and Flash Floods.  NOAATechnical Report

     NESDIS 99.

 

Vicente, G. and R.A. Scofield, 1998: The OperationalGOES Infrared Rainfall Estimator Technique.

     Bull.Amer. Meteor. Soc. 79, 1883-1898.

 

Weisman, M.L. and J.B. Klemp, 1982: The Dependenceof Numerical Simulated Convective Storms on

     VerticalWind Shear and Buoyance. Mon. Wea. Rev., 110,504-520.