Synoptic Climatology of the SRB

SRBEX research in synoptic climatology had five goals in 1994-5:

High Average Discharge

One goal of the research was to see which frequently recurring weather events are consistently associated with high, but not necessarily extreme, discharge. To this end, Yarnal and Frakes (1995a) examined 12 years of daily weather maps from 1978 to 1989 (Yarnal, 1993) using the sequencing methodology of Comrie (1992). These synoptic sequences are somewhat similar to output from mesoscale meteorological models and relate to discharge from individual river basins. Consequently, we analyzed the most frequent two-day and three-day weather-map sequences associated with high average-daily discharge from thirty basins distributed evenly over the SRB. The discharges from these sub-basins operate at three distinct orders of magnitude that correlate with basin size, and are referred to as small, medium and large basins for convenience (Table 2). We analyzed 11 types of storm sequences and found that these storm systems do produce high discharge from the study basins almost every time they occur.


To illustrate, Figure 6 shows the average pressure pattern for the cyclonic conditions with rain synoptic type sequence (RC-RC). The RC-RC is the most frequent of the sequences associated with high average discharge, occurring on average nearly eight times per year. It is most common from late fall through mid-spring, with peak frequencies occurring in February through April; it is infrequent in summer. On Day-1 of this two-day sequence, a low-pressure system moves into the Midwest, while the pressure gradient directs southeasterly flow off the mid-Atlantic coast into the SRB. By Day-0, the low moves east and intensifies, winds over the SRB are now from the east, and precipitation is falling steadily on the basin. The storm system is tapping into the ready supply of Atlantic moisture.

Similar to Yarnal and Draves (1993), we plotted each of the storm tracks associated with the high average-discharge sequences. For instance, the RC-RC storm tracks exhibit a clear northeast ward direction and a distinctive seasonal cycle (Figure 7). Because the circumpolar vortex reaches its southernmost extent in winter, the storms that form in this season tend to come from the southern Great Plains and Gulf of Mexico. Springtime lows originate slightly to the north. The few RC-RC sequences that arise in summer form much farther north and have erratic, slow-moving paths. Early autumn storm tracks look like those of summer, but by late fall originate much farther south, are more intense, and move more quickly.

Similar analysis of the 10 other most frequent weather-map sequences associated with the SRB's high average discharge demonstrates that these synoptic-scale weather systems typically draw their moisture from the mid-Atlantic. This moisture source is quite different to those systems pro ducing extreme discharge in the SRB.

Extreme Discharge

To determine the nature of storm systems that produce extreme discharge events, Frakes and Yarnal (1995c) performed a similar analysis to the one described above. The biggest difference was that weather-map sequences had to occur when discharge was greater than one standard deviation above the lognormal mean. During the 12-year study period, there were 589 days when at least one of the 30 basins experienced extreme discharge and a total of 5383 extreme-discharge basin days. There were 10 two-day or three-day sequences most often associated with these high flows.

For example, the pressure composite of the western side of a high-pressure cell synoptic type sequence (BH-CF) (Figure 8) shows a large Bermuda High over the Atlantic Ocean on Day -1. The strong pressure gradient between the high- and low-pressure centers generates southwesterly winds that advect warm, moist Gulf of Mexico moisture northeastward to the SRB. Typically, a cold-front extends southward from the low. By Day -0, this low has tracked east with the cold front passing over the SRB. As the front approaches, the weakened Bermuda High migrates southeastward, but continues to pump Gulf air into the warm sector ahead of the front. At least during warm months, conditions are stormy, with localized convective precipitation.

The BH-CF tends to be a warm-season phenomenon, with the most occurrences and most frequent extreme discharge during summer months. It does occur infrequently and produce extreme discharge in late winter-early spring; however, in these cases, the high runoff totals are the result of warm-air advection and snow melt on Day-1 not intense precipitation. Of the 234 times the BH-CF sequence occurred during the study period, it induced extreme discharge 120 times in at least one sub-basin of the SRB.

In summary, this analysis demonstrates that the SRB's extreme discharge events usually involve a combination of powerful high-pressure systems, intense pressure gradients, and very warm, moist, southerly air from the Gulf of Mexico. In contrast, high average discharge relies on prolonged low pressure over the basin and easterly moisture flow off the mid-Atlantic coast. Extreme discharge events in the cold season result from the same weather patterns that generate warm-season extreme flows, except that snow melt, rather than localized convection, produces exceptionally high stream levels.

Scales of Response

Through modeling and observation, SRBEX seeks to determine how watersheds of differing scales respond to forcing by weather systems. For the observational component of this objective, Frakes (1995) and Frakes and Yarnal (1995a,b) stratified the discharge from the 30 watersheds by the synoptic type sequences presented above. Composite hydrographs associated with the RC-RC sequence discussed above depict the responses of small, medium, and large basins (Figure 8). Looking only at the heavy, dark line that indicates the average flow from all 10 basins in a size category, the small basins respond strongly on Day-0. Note that discharge is already about 0.25 standard deviations above mean flow on Day-1, suggesting that the first day in the sequence has already saturated the soil by Day-0. Also, the days following this weather-map sequence show gradually decreasing discharge, suggesting raised base flows and the possibility of prolonged wet weather. Individual basins, marked by the fine, dashed lines, have similar response curves, but differ in magnitude. As water funnels into the medium basins from upstream, the hydrographic traces show a stronger response to the RC-RC sequence with a large, rounded peak. This response lags the small basins by a day. The responses of the down stream large basins are as strong as the medium-sized basins, but not stronger, and are delayed two to three days. In contrast to the small and medium-sized basins, large basins have very little deviation around the composite response curve.

It is instructive to compare the RC-RC hydrographs with those of other synoptic type sequences. For example, the cyclonic conditions with rain-cold front sequence (RC-CF), which denotes the classic southwest-to-northeast succession of a warm front followed by a cold front, produces a distinctly different set of basin responses (Figure 9). The small-basin hydrographic traces exhibit a sharp, strong response on Day-0, but stream levels drop off quickly over the next two days. Medium-sized basins also have a sharp, but slightly weaker response a half-day later. The large basins take two days to achieve peak response, and flow levels do not drop off rapidly after the event.

Thus, this part of the research confirms a well-known principle in basin hydrology: small basins respond the quickest to runoff events, while the response times of larger basins lag according to their distance downstream from the headwaters. However, what is unique to this research is the conclusion that this lag time varies with the storm trajectory and type.

Representative Hydroclimatic Events

Nested-modeling efforts, such as SRBEX, assume that there is a strong causal relationship between the atmospheric circulation and the hydrologic response of the basin. Many of the simulations linking the atmosphere and basin hydrology rely on observed events believed to be typical. Yet no methodology has been developed to determine either the representativeness of the circulation patterns producing a significant hydrologic response, or the representativeness of the hydrologic response to a given storm system. Thus, to determine which circulation patterns and their hydrologic responses are representative, Yarnal and Frakes (1995b) scrutinized the results presented above.

As suggested by the differences between the RC-RC and RC-CF sequences, each weather-map sequence produces a unique hydrographic response. In any one size category and assuming somewhat similar geology, vegetation, and soils, only those basins that have suffered extreme land-surface modification or flow control show radically differing hydrographs. This is an extremely important finding because it suggests that, in a general region, almost any basin with good measurements is "representative" of its scale.

This work also demonstrates that hydrologic and meteorological models must be able to simulate the flow from many different precipitation patterns over time and space. If a hydrologic model reproduces the hydrographic trace of a generic, average precipitation or snow melt event, it does not necessarily mean that the model is a success; it must be able to generate lifelike responses to many kinds of weather events. Similarly, mesoscale meteorological models must be able to simulate the precipitation from a wide range of meteorological scenarios. Thus, the results suggest that synoptic climatology is a useful tool for defining model scenarios and for model verification in global-change experiments.

Monitoring Large-Scale Hydroclimatic Variation and Future Work

The fifth goal of the synoptic climatological research was to develop ways to monitor large-scale climatic variation over the SRB. Yarnal and Frick (1995) are creating a catalog of daily eigenvector-based synoptic map patterns, plus indices based on these patterns (Yarnal, 1993) for the period 1946-1994. Quarterly updates of the catalog will be available from the Northeast Regional Climate Center in hard copy and over the Internet starting in the last quarter of 1995. These data will be integral to developing long-range forecasts of basin climate and hydrology.

From this starting point, the next phase of the synoptic climatological work will focus on long-term, large-scale climate-runoff relationships in the SRB. To date, the research has concentrated on relationships between daily synoptic-scale circulation and discharge at various basin scales. The new emphasis will be on translation among both spatial and temporal scales. Specifically, the research will involve: Understanding these spatial and temporal relationships is intrinsically important to basic science. Furthermore, developing knowledge of these associations provides a baseline for guiding and evaluating the nested-model experiments of SRBEX. This work is especially important to linking the GCM and mesoscale meteorological model and projecting future basin responses from large-scale model and satellite-observed data.

Penn State/NCAR Mesoscale Model

The Penn State/NCAR Mesoscale Model (MM) is a versatile three-dimensional, limited-area meteorological numerical model. The 5th generation of the MM (MM5) has been chosen for SRBEX applications because it is a state-of-the-science model. MM5 is nonhydrostatic, it allows for multiple movable nested domains, and it has a sophisticated atmospheric hydrologic cycle module, an improved atmospheric radiation scheme, and efficient numerics to make longer-range simulations more efficient. It can be initialized with observed or synthetic atmospheric data and can be applied to any region on earth. MM5 has been tested on a number of meteorological cases over a wide range of scales where the emphasis has been on the hydrologic cycle and the surface forcing. All these features make MM5 an excellent choice for studying regional climate changes with emphasis on the water cycle.

Version 1 of MM5 was officially released for public use in February 1994. The entire MM5 system (including pre- and post-processing programs) was transferred to the ESSC CRAY YMP, and the necessary changes and tests to run those programs on that platform were made.

During 1994 a modified version of the Biosphere-Atmosphere Transfer Scheme (BATS) (Lakhtakia and Warner, 1994) was introduced into MM5, so that the latter can be used as part of the SRBEX suite of models. Much emphasis has been put on the improvement of some of the variables/parameters within BATS that require initialization/specification. BATS requires the initialization of the soil water (SW) content and the specification of some surface characteristics. At present, MM5 is the model being used to develop and test the linkages with the SHM and Terrestrial Hydrology Model (THM) described below (and shown in Figure 3). In the near future, MM5 will replace the RegCM2 (based on MM4) as the model coupled to the GCM. The MM5 multiple-nesting capability will provide dynamically consistent "downscaling" capabilities required to link the large scale of the GCM to the local scale of the THM.

Table of Contents Previous Section Next Section Download Report
Graphics and HTML text markup by Phil Kolb
Comments to: R. White, raw@essc.psu.edu