摘要: |
Sustainable construction and maintenance of surface transportation systems requires accurate hydrologic design. A critical element of hydrologic design is the estimation of flood discharges that bridges and culverts must convey and their abutments must withstand. Such flows are produced by the combination of rainfall with a prescribed frequency of occurrence and the conversion of the rainfall into runoff and, ultimately, streamflow. In some cases, the streamflow can be estimated using statistical methods, but for many ungauged basins, the flows are estimated by rainfall-runoff modeling. An important and long-standing problem for the Rocky Mountain Region is that traditional rainfall-runoff modeling methods appear to significantly overestimate major floods based on comparisons to paleoflood evidence and regional peak streamflow statistics. The Colorado Water Conservation Board (CWCB), Colorado Division of Water Resources (DWR), and State of New Mexico are currently conducting a $1.5 million study to develop improved estimates of extreme rainfall for the two state region. DWR has also been working to improve flood hydrology methods for the Rocky Mountain region. Solving this problem will align with these ongoing efforts and allow higher confidence in bridge and culvert design, more efficient allocation of transportation repair and replacement funds, and the possibility of streamlined design guidelines.
Traditional flood hydrology methods utilize low infiltration rates to model flood runoff solely by an infiltration-excess mechanism. By this mechanism, runoff occurs when the rainfall intensity exceeds the infiltration capacity of the soil. However, forested basins typically have soils with high infiltration capacities that produce little infiltration-excess runoff (Dunne and Leopold, 1978; MacDonald and Stednick, 2003). Furthermore, the bedrock geology of many mountain basins leads to coarse soils with high infiltration rates. Recent advances by DWR indicate that flood runoff in mountain basins might be controlled by a saturation-excess mechanism (DWR, 2014 and 2015). Saturation-excess runoff can occur when a relatively shallow soil is underlain by a layer with much lower permeability (usually bedrock), which is a relatively common situation in Rocky Mountain basins. Rainfall rates that are less than the infiltration capacity can still produce runoff if the storm continues long enough to saturate the thin soil layer. Such lowintensity events are expected to be more important at higher elevations where strong convective storms are less common (Grimm et al., 1995). A recent but preliminary examination of the Gross Reservoir basin (South Boulder Creek) that was performed by Colorado Dam Safety suggests that saturation-excess runoff might be important for extreme precipitation events (Perry and Franz, 2017). For the September 2013 storm that produced widespread flooding along the Colorado Front Range, the rainfall rate in the Gross Reservoir basin never exceeded 1.2 in/hr, but the storm continued for about 6 days. During that period, two peaks in rainfall intensity occurred approximately one day apart. Although the first peak had a higher rainfall intensity, the second peak produced much more runoff (Fig. 1). This behavior is not consistent with infiltrationexcess runoff, which would produce higher runoff rates for higher rainfall rates, but it is consistent with saturation-excess runoff, which depends more on the accumulated depth of rainfall.
Similarly, a study sponsored by the Colorado Department of Transportation (CDOT) calibrated a rainfall-runoff model to 10-day September 2013 flows in the upper Big Thompson River basin and then used frequency rainfall estimates in the model to estimate frequency flows for bridge and culvert design. Their calibration efforts indicate high rainfall losses and almost no runoff until 5 days into the 10-day period, followed by a sudden change to minimal losses and high runoff (Jacobs, Inc., 2014). Although a physical explanation was not provided, these results are consistent with saturation-excess runoff.
Finally, both the Colorado Dam Safety and CDOT studies had difficulty reproducing the observed hydrograph recessions for historic long-duration events in mountain basins. This problem appears to be associated with lateral subsurface flow (i.e. interflow) and may be consistent with a saturation-excess model where high volumes of water are temporarily stored in relatively coarse-grained shallow soils. |