A Forecast Overview of the 26 January 1996 Blizzard Across Eastern Iowa and Northwest Illinois Using PCGRIDDS and Non-Standard Forecast Techniques


Todd E. Holsten and Tim Hendricks
National Weather Service
Davenport, Iowa



During the late morning and early afternoon of 26 January 1996, heavy snow fell in a 50 mile wide band across eastern Iowa and extreme northwest Illinois. Accumulations ranged from 1 to 13 inches in the National Weather Service Davenport, Iowa (NWSO DVN) county warning area (Figure 1). Most of the snowfall occurred in less than 8 hours. Many locations along the Mississippi River reported lightning and thunder with the snow and snowfall rates as high as 3 inches per hour.

Figure 1 . Observed snowfall totals across eastern Iowa and northwest Illinois from the 26 January 1996 blizzard.

In this study, a description of the synoptic-scale environment that caused the heavy snow to develop will be included with an emphasis given to the analysis of the NGM numerical model guidance using the PC-Gridded Interactive Diagnostic and Display System (PCGRIDDS) software. The analysis and role of coupled jet streaks will be assessed or analyzed using vertical motion fields, Q-vectors, and ageostrophic circulations to illustrate the potential for heavy snow across eastern Iowa and northwest Illinois. An examination of a tropopause fold and its possible role in rapid cyclogenesis is included. Additionally, the potential for conditional symmetric instability and banded convective mesoscale precipitation will be examined. Finally, a comparison of quantitative snow forecasting techniques will be presented.


At 0000 UTC 26 January 1996, a potent 500 mb shortwave trough was located across southern Utah with a developing surface low (as inferred from mean sea-level pressure) (MSLP), in northeast New Mexico. At 1200 UTC, the 500 mb negatively-tilted trough (open wave) was forming over western Nebraska and Kansas with a 1006 mb surface low located near Tulsa, Oklahoma. A warm front stretched from this low into central Missouri. By 0000 UTC 27 January, the 500 mb trough was centered over southeast Iowa with the surface low in southwest Wisconsin. During this latter 12-hour period, the surface low deepened explosively with the central pressure falling 16 mb. The surface low actually tracked almost due north from central Missouri to eastern Iowa during the morning hours as it rapidly intensified.

An area of heavy snow rapidly expanded across eastern Iowa and extreme northwest Illinois in response to very strong warm air advection between 850 and 700 mb. This was driven by a low-level jet of 40 to 50 knots from Louisiana into Iowa. During the height of the storm at 1700 UTC 26 January, Mesoscale Analysis Prediction System [MAPS] (Benjamin et al. 1994) surface analysis (Figure 2) showed an intense surface low over northern Missouri with three-hour pressure fall "bulls eyed" over southeast Iowa. Six-hour pressure falls (14-20 UTC 26 January 1996, not shown) of 14 mb occurred across eastern Iowa as the surface low rapidly intensified.

Figure 2.

Light snow began falling at NWSO DVN at 1200 UTC 26 January and quickly became heavy by 1300 UTC. KDVN WSR-88D radar showed distinct higher reflectivity bands moving north along the Mississippi River between 1300 and 2100 UTC 26 January. Thundersnow with snowfall rates of 1 to 3 inches per hour were reported in many locations along and west of the Mississippi River during this period. Storm total snowfall ranged from 6 inches along the Mississippi River to 13 inches in northeast Iowa. A distinct cut-off to the snow was evident, with areas east of the Mississippi River in northwest Illinois receiving much less. For example, Davenport, Iowa received 10.7 inches, while locations 20 miles east only received an inch (Figure 1). The snow ended across most of eastern Iowa and extreme northwest Illinois during the early afternoon hours as the system lifted northeast and the mid-level dry slot moved across the region. Central and northeast Iowa were closer to the mid-level deformation zone; therefore, the event was of longer duration there with the snow continuing into early evening.


A. Jet Streak and Ageostrophic Circulation

Several studies have documented the importance of jet streak interactions to the production of enhanced vertical motions and heavy snowfall. Uccellini and Kocin (1990) showed that upward vertical motion is enhanced by the merger (coupling) of the ascending branches of the two separate jet streaks, and this often accompanies heavy snowfall along the East Coast. Hakim and Uccellini (1992) documented jet streak coupling for a snow band across the Northern Plains, while Shea and Przybylinski (1993) did likewise for a snowstorm in Missouri. The lift occurs in the ascending branch of the thermally direct transverse vertical circulation within the entrance region of the first jet, and the thermally indirect transverse circulation within the exit region of the second jet streak.

The 12-h NGM forecast valid 1200 UTC 26 January revealed a well-defined 300 mb jet streak across the Great Lakes and another across Oklahoma (Figure 3). The entrance region of the northern jet and the exit region of the southern jet appeared coupled over the middle Mississippi River Valley. As a result of the strong along-stream variation in wind speed, the coupled ageostrophic wind response was significant in enhancing the upward vertical motion.

Figure 3. NGM 12-h forecast of 300 mb wind and isotachs (kt), valid at 1200 UTC 26 January 1996.

A spatial-height cross-section (Figure 4), derived from the 12-h NGM forecast valid 1200 UTC 26 January, depicts the transverse circulation of ageostrophic wind. The direct thermal circulation (D), and the pronounced deep-tropospheric upward motion are clearly defined between the two jet streaks. The location and timing of this enhanced ascent are coincident with an increase in precipitation intensity across eastern Iowa and northwest Illinois. In addition, the exit region of the second jet streak likely intensified the southerly low-level jet, resulting in enhanced low-level convergence and ascent across eastern Iowa and northwest Illinois (Uccellini and Kocin 1990). A spatial-height cross-section (Figure 5), from the 18-h NGM forecast valid 1800 UTC 26 January, shows a pronounced divergence/convergence couplet over eastern Iowa and northwest Illinois, coincident within the area of strong upward vertical motion.

Figure 4. NGM 12-h forecast of ageostrophic circulation resulting from the coupling of two separate jet streaks, valid at 1200 UTC 26 January 1996. The cross-section extends from El Paso, Texas (left) to Montreal, Quebec (right). The thermally direct circulation is marked by the letter "D". Note the deep-tropospheric upward motion (µb/s, dashed lines) between the two jet streaks (kts * 10, solid lines).

Figure 5. NGM 18-h forecast valid at 18 UTC 26 January 1996. The cross-section is denoted by the solid straight line in Figure 3 and extends from El Paso, Texas (left) to Montreal, Quebec (right). Note the pronounced divergence (s-1, solid lines) /convergence (s-1, dashed) couplet from southern Kansas into western Illinois as denoted between the arrows.

B. Q-vector Diagnosis

Q-vectors are a useful tool for diagnosing synoptic-scale areas of vertical motion within the dynamical constraints of quasi-geostrophic (QG) theory. A good review of the operational use of QG theory and Q-vectors can be found in Barnes and Colman (1993).

Figure 6 depicts the 18-h NGM forecast of 850-300 mb layer Q-vector and Q-vector divergence. Note how strong forcing for ascent, as implied by the areas of Q-vector convergence or negative divergence (dashed), is implied in the area previously shown to be experiencing upward vertical motion due to the ageostrophic secondary circulations from the two jet streaks. The area of implied upward vertical motion, as described by the layer Q-vectors, accurately forecast the beginning and ending times of the snow.

Figure 6. NGM 18-h forecast 850-300 mb layer mean Q-vectors and divergence of Q (10-19 m/Pa m s, dashed contours indicate convergence or negative divergence).

C. Potential Vorticity and Surface Cyclogenesis

Temperature changes high in the troposphere or low in the stratosphere can have a greater impact on lower-level height tendencies than changes in the lower to middle troposphere (Hirshberg and Fritsch 1991a, b). Thus, high-level temperature advection (associated with tropopause undulations) is a very powerful forcing mechanism on the lower-level height fields. Tropopause undulations can act synergistically with lower tropospheric processes to initiate, maintain, and enhance the development of extratropical cyclones.

Tropopause undulations are characterized by high levels of potential vorticity ( > 1*10-6 K/mb s) from the stratosphere and are more stable, warmer, and drier relative to nearby tropospheric air. According to Holiway and Smith (1995), with the tropopause undulation, stratospheric air lowers into the trough axis of the wave into the troposphere and warms adiabatically. This warmer air at the 200 and 300 mb levels then aids in the development of a surface cyclone by decreasing the density of the air column that is applying pressure from above. With less weight above the cyclone, the storm has the opportunity to intensify as surface pressure decreases at its center. These tropopause undulations can have a profound effect on developing low pressure systems as shown by Hirschberg and Fritsch (1991a,b) and Holiway and Smith (1995).

The NGM 18-h forecast valid at 18 UTC 26 January (Figure 7), shows this potential vorticity field in the 500 to 300 mb layer. Note the high values of potential vorticity across Kansas. At this time, the MSLP low was rapidly intensifying across northeast Missouri (Figure 2) possibly in response to the tropopause fold that is depicted in a cross-section (Figure 8). The NGM model run from 0000 UTC 26 January 1996 predicted the MSLP low to develop slowly and track from eastern Oklahoma to northern Illinois through 24 hours. Analyzing the potential vorticity fields can provide insight into the possibility for rapid development of surface cyclones, and thus the potential of blizzard conditions.

Figure 7. NGM 18-h forecast layer mean 500-300 mb Potential Vorticity (10-6K/mb s, solid lines) valid at 18 UTC 26 January 1996.

Figure 8. NGM 18-h forecast cross-section valid at 18 UTC 26 January 1996. Potential vorticity (10-6K/mb s, solid lines) and equivalent potential temperature (K, dashed lines). The cross-section extends from Medford, Oregon (left) to Atlanta, Georgia (right), which is denoted by the solid straight line in Figure 13.

D. CSI - Conditional Symmetric Instability

Thundersnow reports were widespread across eastern Iowa and northwestern Illinois during the morning. At 1819 UTC 26 January 1996, KDVN Doppler radar 0.5 degree composite reflectivity (Figure 10) indicated several narrow bands of 35-45 dBZ reflectivity echoes across eastern Iowa and northwest Illinois, similar to those described by Bennets and Sharp (1982).

Angular Momentum (Mg) and Equivalent Potential Temperature (e) surfaces were constructed via PCGRIDDS using a cross-section taken normal to the 1000-500 mb thickness.

According to Snook (1992), CSI exists in saturated regions where the slope of Mg is equal or less than that of e. Moore and Lambert (1993) offer an empirical value of 80 percent relative humidity in the area of CSI. Investigation of the 00 UTC 26 January 1996 NGM 18-h forecast, using a cross-section drawn perpendicular to the thermal wind, indicated conditions could be favorable for CSI over eastern Iowa and northwest Illinois, although is it questionable how well CSI can be resolved given the relatively course grid scale of the NGM.

The 1200 UTC 26 January Davenport, Iowa sounding (DVN) (Figure 9) indicated air parcels lifted from the top of the inversion were not convectively stable with a 700-500mb layer lapse rate of -6.5 degrees C/km. Thus, upright moist elevated convection was a more likely possibility and was supported by satellite imagery.

Figure 9

Figure 10

Recognition of CSI, or elevated instability in general, is meteorologically important because periods of heavier snowfall (perhaps with thunder), and higher accumulations than may otherwise have been expected could occur. This would obviously impact public forecasts, but could also impact QPF forecasts, which are input into hydrologic models, and visibility forecasts for the aviation community.


Strong synoptic-scale upward vertical motion of moist air produced heavy precipitation across the entire middle and upper Mississippi River Valley. Much of this lift was accomplished through strong warm air advection. One non-standard and three commonly used QPF forecast methods were applied to the 00Z 26 January 1996 NGM forecast model run.

A. Cloud Microphysics

Heavy snowfall rates are typically associated with large snowflakes consisting of dendritic crystals. Auer and White (1982) suggested that heavy snow can occur in events when the level of maximum vertical motion is near the level of maximum growth rate of dendritic crystals, which typically occurs at temperatures of -13 to -17 degrees C. The period of heavy snow in eastern Iowa and northwest Illinois correlated with the time period when the area of maximum upward vertical motion occurred within the -13 to -17 degrees C temperature range, as indicated by the derived time-height cross-section for Davenport, Iowa (Figure 11; from 1500 to 2100 UTC 26 January).

Figure 11. Time-height cross-section for Davenport, Iowa from the 00 UTC 26 January 1996 NGM. Note the level of maximum vertical motion (µb/s, dashed lines) in the maximum growth rate temperature level (degrees C, solid lines).

B. Garcia Method

The 6-h through 18-h 288 K isentropic surface, based on the 0000 UTC 26 January 1996 NGM model run, indicated a tight gradient of pressure and mixing ratio across eastern Iowa and northwest Illinois, with strong cross-isobaric and cross-isohume flow from higher-to-lower pressure and mixing ratio (Figure 12). Thus, pronounced adiabatic upward motion (50 mb of lift) and moisture advection (condensation pressure deficits were less than 5 mb) were present in eastern Iowa and northwest Illinois along the warm conveyer belt.

Figure 12

A good correlation existed between isentropic mixing ratios and the maximum amount of snowfall during the ensuing 12 hours (Garcia, 1994). According to Garcia, the desired isentropic surface should intersect the 700-750 mb layer over the area of concern, which in this case was eastern Iowa and northwest Illinois. The isentropic surface 288 K was chosen as the "best fit."

An effective mixing ratio value for the 12-hour period can be calculated by averaging the mixing ratio over the area of concern and the highest mixing ratio that could be advected into this area. The average mixing ratio value provides the basis for the maximum snowfall forecast. Multiplying the average wind speed by the time period of 12 hours will give the approximate distance to the point where the highest mixing ratio can be obtained.

In this event, the average wind speed was determined to be 32 kts and the estimated distance from the area of concern was 445 mi (Ft. Smith, AR). The maximum mixing ratio was 4 g/kg. Taking the average of the mixing ratios (1.5 and 4.0), yielded an effective mixing ratio of 2.75 g/kg. According to Garcia's scale (2 to 1 ratio), this translates to a forecast of 4-6 inches of snow.

C. Cook Method

The Cook 1980 method relates the magnitude of 200 mb warm air advection to snowfall amounts. Approximately one-half of the value of the indicated warm air advection (C) at 200 mb is used for the amount in inches, assuming warm air advection is occurring at 700 mb. If cold air advection is occurring at 700 mb, one-quarter of the value is used. This technique assumes the following: (a) existence of a surface low, (b) no northwest flow or areas of strong confluent flow at 200 mb, (c) winter season event, (d) system is well developed. The Cook method indicated 8° to 10 degrees C temperature advection at 200 mb through 24 hours, and the potential of 4 to 5 inches of snowfall across eastern Iowa and northwest Illinois (Figure 13).

Figure 13. NGM 24-h forecast 200 mb warm air advection from 06 UTC 26 January to 06 UTC 27 January 1996. Positive values (°C/24hr, solid lines) denote warm air advection.

D. Magic Chart

The Magic Chart (Sangster and Jagler 1985) was developed as a guide to help forecasters estimate snowfall potential using 12 hour 700 mb net vertical displacement (NVD) from the trajectory model (Reap 1990). The heaviest snowfall is likely to occur in the thermal region between -3 and -5 degrees C at 850 mb with the greatest NVD. The assumptions for applying this method are similar to the assumptions of the Cook method in that well developed or developing upper tropospheric and surface systems are necessary, and at least 90 percent relative humidity (RH) in the 1000-500 mb layer is required. A limitation of this method is the temperature range between -3 and -5 degrees C, which makes the technique inappropriate in events involving an arctic air mass. Although not shown, the Magic Chart criteria were met over southeast Iowa to northwest Illinois with 24-h forecast 12-h net vertical displacements of 60 to 80 mb indicating the potential for 6 to 8 inches of snow across parts of Iowa and Illinois.


NGM 24-h QPF forecast indicated a broad area of ¼ to ½ inch of stratiform precipitation across most of Iowa and northwestern Illinois (Figure 14). The NGM was not forecasting any convective precipitation. Surface temperatures in eastern Iowa and northwestern Illinois during the morning of 26 January 1996 were in the mid 20s. Assuming a snow-liquid equivalent ratio of 10 to 1, 3 to 6 inches of snow could be expected. This agreed with the previously discussed QPF techniques and pointed to a maximum of 6 inches of snow expected in the heaviest snow band.

Figure 14. NGM 24-h QPF from 0000 UTC 26 January to 0000 UTC 27 January (contoured every .05 inch).

In summary, the various QPF techniques along with the NGM 24-h QPF forecast indicated the potential of 3 to 6 inches of snowfall over eastern Iowa and northwest Illinois. This was about half the snowfall that actually occurred. This gross underestimation was likely due to the models underestimating the intensity of the storm system. Most of the actual QPF was convective rather than stratiform.


Forecasters expected the heaviest snow to fall from Burlington, Iowa to Rockford, Illinois. This was based on the forecast track of the system which was further east due to the NGM's slower development of the system than what actually transpired. The heaviest snow actually fell from south central Iowa through northeast Iowa and into Wisconsin in association with the deformation zone. Parts of Iowa were blanketed under 10-14 inches of snow and blizzard conditions as winds increased rapidly to 45 miles per hour in response to the strong pressure gradient between the rapidly intensifying surface low in Missouri and the arctic high building in the Dakotas. A 50-mile wide band of heavy snow (6 to 10 inches) did fall along the Mississippi River in less than 8 hours during the late morning and early afternoon on 26 January 1996. This snowband was associated with the strong warm air advection that was occurring north of the warm front on the nose of the low-level jet, and was enhanced due to the coupled jet streak interaction and intense dynamical forcing for rising motion. The use of the potential vorticity fields provided forecasters with a hint that this system could intensify more rapidly than what was forecast, and that blizzard conditions could possibly occur.

A comparison of snow forecasting techniques currently in operational use showed the following. The Cook method indicated 4 to 5 inches of snow from 0000 UTC 26 January to 0000 UTC 27 January across Iowa. The Magic Chart indicated a narrow swath of heavy snow across southeastern Iowa into northwest Illinois of 6 to 8 inches. Finally, the Garcia method indicated 4 to 6 inches was likely across eastern Iowa and northwest Illinois. Actual snowfall amounts were much higher than any of the quantitative forecast methods indicated. This was likely due in part to the NGM under developing the system and underplaying the convection. Convection, whether slant wise or upright, likely played a role in enhancing snow amounts above what any of the QPF methods indicated.


The authors would like to thank Ray Wolf (SOO - NWSO Davenport, Iowa) and Ed Berry (Central Region Scientific Services Division) for their guidance during the preparation of the paper and their comments and suggestions throughout the review processes.


Auer, A.H., Jr., 1987: An aid to forecasting heavy snowfall episodes. Nat. Wea. Dig., 12, 11-14.

Barnes, S.L., and Colman, B.R., 1993: Quasigeostrophic diagnosis of cyclogenesis associated with a cutoff extratropical cyclone-The Christmas 1987 storm. Mon. Wea. Rev., 21, 1613-1634.

Benjamin, S.G., Brundage, K.J., and Morone, L.L., 1994: The Rapid Update Cycle Part 1: Analysis/Model Description. NWS Technical Procedures Bulletin, 416, 1-16. U.S. Dept. of Commerce, NOAA, Silver Spring, MD.

Bennets, D.A., and J.C. Sharp, 1982: The relevance of conditional symmetric instability to the prediction of mesoscale frontal rainbands, Quar. J. Roy. Meteor, Soc., 108, 595-602.

Cook, B.J., 1980: A snow index using 200-mb warm advection. Natl. Wea. Dig., 5, 29-40.

Garcia, C., Jr., 1994: Forecasting snowfall using mixing ratios on an isentropic surface - An empirical study. NOAA Technical Memorandum NWS CR-105. DOC, NOAA, NWS, Central Region, Scientific Services Division, Kansas City, MO, 28 pp.

Hakim, G.J., and L.W. Uccellini, 1992: Diagnosing coupled jet streak circulations for a Northern Plains snow band from the operational Nested-Grid Model. Wea. Forecasting, 7, 26-48.

Holiway, C.E., and D.R. Smith, 1995: The role of tropopause undulations in the development of the "Blizzard of '93" (12-15 March 1993), Preprints, 14th Conf. Wea. Analysis and Forecasting, Dallas, TX., AMS, 80-84.

Hirschberg, P. A., and J.M. Fritsch, 1991a: Tropopause undulations and the development of extratropical cyclones. Part I: Overview and observations from a cyclone event. Mon. Wea. Rev., 119, 496-517.

____________, and ____________, 1991b: Tropopause undulations and the development of extratropical cyclones. Part II: Diagnostic analysis and conceptual model. Mon. Wea. Rev., 119, 518-550.

Moore, J.T., and T.E. Lambert, 1993: The use of equivalent potential vorticity to diagnose regions of conditional symmetric instability, Wea. And Forecasting, 8, 301-308.

Reap, R.M., 1990: Forecasting snow amounts from net vertical displacements. NWS Technical Procedures Bulletin, 389, 1-4. U.S. Dept. of Commerce, NOAA.

Sangster, W.E., and E. Jagler, 1985: The (7WG, 8WT) "Magic" Chart, CR Technical Attachment 85-1. DOC, NOAA, NWS, Central Region, Scientific Services Division, Kansas City, MO, 5 pp.

Shea, T.J., and R.W. Przybylinski, 1993: Assessing vertical motion fields in a winter storm using PCGRIDDS. Preprints, 13th Conf. Wea. Analysis and Forecasting. Vienna, VA., AMS, 10-14.

Snook, J.S., 1992: Current techniques for real-time evaluation of conditional symmetric instability, Wea. Forecasting, 7, 430-439.

Uccellini, L.W. and P.J. Kocin, 1990: The interaction of jet streak circulations during heavy snow events along the east coast of the United States. Wea. Forecasting, 2, 298-308.


USA.gov is the U.S. government's official web portal to all federal, state and local government web resources and services.