WESTERN REGION TECHNICAL ATTACHMENT
NO. 03-10
November 26, 2003

The Campo Bonito Wash Flash Flood of August 14, 2003:
A Heavy Rain Event on a Recent Burn

Mike Schaffner, NOAA/NWS Weather Forecast Office, Tucson, AZ

Introduction

The Santa Catalina Mountains rise up from the valleys of the Tucson metro area (an elevation of about 3,500 feet, to several summits over 9,000 feet). The area is mostly within the Coronado National Forest. Several forest fires have impacted the Santa Catalinas since 2002. The Bullock Fire burned 30,563 acres in May and June, 2002. The Oracle Hill fire consumed 2,432 acres in July, 2002. The Aspen Fire burned over 84,000 acres in June and July, 2003.

During the 2002 monsoon season, Campo Bonito residents reported several high flows near bankfull of approximately 350 cfs. Although this was an increase in high flows from pre-Oracle Hill Fire conditions it was not a cause for alarm from local residents.

The Campo Bonito basin is two square miles (Figure 1) and includes about nine private in-holdings within the Coronado National Forest. Sixty percent of the basin is currently burned: 0.45 square miles being from the Oracle Hills Fire and 0.75 square miles from the Aspen Fire. Within the basin, the Aspen Fire burned approximately one-quarter high severity burn, one-half moderate severity burn, and one-quarter low severity burn while the Oracle Hill Fire was dominated by high severity burn (Figure 2 and 3). An early stage of recovery was noticed by the presence of new roots sprouting from the bases of burnt oaks within the Oracle Hill Fire burn area.

Event Overview

On the afternoon of Thursday, August 14, 2003, a number of storm cells converged and moved west from the Galiuro Mountains over the San Pedro Valley, near San Manuel, and through the northern Santa Catalina Mountains near Oracle in Pinal County (Figure 4). Storm motion was to the west-southwest or generally moving up basin.

KEMX Tucson radar estimated 1.60 to 1.75 inches of precipitation fell between 4:00 and 4:25 PM MST (Figure 5). A spotter in Oracle reported 1.8 inches over the same time period. A Pima County Flood Control ALERT gage, just to the west on Oracle Ridge, reported 1.54 inches between 4:07 and 4:36 PM MST. While spotters also reported small hail with this storm, radar estimates approximately match rain gage data in the area. A rainfall of 1.59 inches in thirty minutes for the Oracle area (ORACLE 2 SE, ARIZONA (02-6119) 32.60.25ºN 110.73.44ºW 4399 feet) amounts to a 25 year rainfall according to the NOAA Atlas 14 http://hdsc.nws.noaa.gov/hdsc/pfds/sa/az_pfds.html. This rainfall resulted in an 8 to 12 foot wall of water in Campo Bonito Wash. Three homes and several other properties were damaged. A long-time resident of the Oracle area was killed when he was swept downstream from his residence. His body was located a half mile downstream the following morning.

While no one else was in the area to witness the event, a coworker reported talking to the deceased at 5:00 PM. This provided a minimum of 5 hours lead time on the flash flood watch issued at 12:05 PM and 34 minutes for the flash flood warning issued at 4:26 PM.

Flood Survey Findings

A flood survey was conducted within 24 hours of the event. The NWS Tucson Hydrologist and WCM, county officials, and state division of emergency management were present. Highwater marks were approximately 5 feet high on the residence of the deceased or about 9 feet high from the pre-flood channel bottom of Campo Bonito Wash. Pre-flood channel depth was based on conversations with neighbors. Campo Bonito Wash makes a greater than 45 degree turn near the residence. A large piling of debris can be seen up against the channel side of residence (Figure 6). A culvert and wooden beam bridge belonging to a neighbor 100 yards downstream was plugged and may have caused much of the channel sedimentation near the home (Figure 7). Further evidence of man-made channel modification was noted several hundred yards downstream including walls that constricted channel flow.

Slope-Area Measurement Findings

During a flood it is often impossible or impractical to measure peak discharges. This may be due to road closures, stressed resources during the event, and sites being ungaged. A slope-area measurement is a common indirect technique used by the USGS to estimate discharge after a flood event. The slope-area measurement makes use of the energy equation for computing streamflow. The energy equation makes use of the physical characteristics of the channel, water-surface elevations at time of peak discharge, and channel roughness coefficients. The data required for a slope-area measurement is obtained in a field survey of a reach of the channel. The survey includes the elevation and location of high-water marks corresponding to the peak stage, cross sections of the channel along the reach, and selection of a roughness coefficient.

The NWS Tucson hydrologist worked with the USGS Tucson water resources field office to make a slope-area measurement for the event. Two slope-area reaches were flagged and surveyed within less than a week of the event (Figure 8, 9, and 10). The upstream reach has an area of 1.25 square miles and the downstream reach an area of 2.0 square miles. The majority of the area burned is within the drainage area for the upstream reach. See Table 1 for slope area results. Bill Reed from CBRFC also calculated pre and post-burn return flows for the site (Appendix A). Post-burn flows were now 2.5 times pre-burn flows for an equivalent rainfall event As a result; a pre-burn 100 year return flow now has a return interval of less than 25 years.

Conclusions

The combined influence of the Aspen and Oracle Hill Fires resulted in significantly increased runoff. Post-burn runoff was about 2.5 times pre-burn runoff for the area of interest. Based on the USFS (Burn Area Emergency Reaction (BAER) Meeting, July 7, 2003) and literature on the subject, it will take three to five years for the basin to recover to pre-burn conditions. The lack of debris dams, as seen by the USFS (Bonito Canyon USFS internal memo, August 21, 2003) and NWS/USGS, indicates that debris was not a significant issue in this particular flood event. Channel constrictions and plugging of local residence bridge culverts downstream of the residence may have contributed to storm damage. However the magnitude and velocity of the peak crest flow and the channel geometry adjacent to the residence seems to be the major cause of the damage.

Forecasters anticipated flash flood potential was very high due to recent fires. The Bullock Fire was the first exposure most forecasters had at WFO Tucson with post-burn flash flood forecasting. The Bullock Fire was on the sparsely populated backside of the Santa Catalinas and thus allowed our office to learn in a safe environment. We were fortunate that the backside of mountain did not experience any reported flooding. This was likely due to the low population density and the lack of rain and stream gages. An unpublished report entitled "Anticipated Flash Flood Impacts from the Aspen Fire" was also written by the NWS Tucson hydrologist. This report was distributed to the staff prior to the start of the 2003 monsoon season. Field work, such as slope-area measurements and flood surveys, revealed important details of post-burn flash floods for operational forecasters dealing with burn-related flash floods. An unpublished report entitled "Canada del Oro (CDO) Wash Flash Flood Survey of the August 1, 2003 Event" was distributed to the staff by the NWS Tucson hydrologist over a week prior to the August 14th event on Campo Bonito Wash. The CDO adjoins the Campo Bonito basin on its west side. A combination of situational awareness, and research both before and during the 2003 monsoon season, contributed to the extended lead times on both the watch and warning for the area.

Acknowledgements

Pinal County emergency manager, Pete Weaver, and department of public works project engineer John Rotter, facilitated the initial flood survey on Friday, August 15, 2003. In addition, John assisted in locating slope area reaches on August 18, 2003. The USGS responded quickly to the situation on Campo Bonito Wash. Within less than four days following the event, the author and Chris Smith (USGS AZ water resources division data chief) were able to walk much of the wash looking for slope area reaches. Two days following that, the USGS had a crew of three out to do the survey. This was a significant expenditure of resources for the USGS. Chris also reviewed this TA. Bill Reed, Colorado Basin River Forecast Center (CBRFC) senior hydrologist and lower Colorado River focal point, calculated pre and post burn return flows. Bill has worked with previous burn areas including the Rodeo-Chediski in central Arizona of 2002. Bill also reviewed this TA and provided comments. The Tucson WFO SOO also reviewed this TA and provided direction. Salek Shafiqullah, a hydrologist with Coronado National Forest, provided the merged GIS burn severity map.