Its been a soggy spring, just like the wet spring of 1995. The ground is soaked and the ditches are running high with snowmelt. One day in mid-June a storm builds up over the foothills and cuts loose. Flood sirens ring out alerting residents of a flood. We only have 45 minutes warning. Manhattan Middle School just let out and the students are waiting to be picked up. A senior couple are in their car driving back to their Boulder Creek Commons townhome. A resident on Cimmarron looks in their backyard and sees Dry Creek Ditch topping out and lapping into their backyard as it backs up. The detention pond at the northwest corner of Boulder Creek Commons quickly fills with the storm flow from the streets. The ditch and storm runoff overflow rushes through Keewaydin Park and surrounds the school. Distraught parents hurrying to pick up their children are trapped in their cars. The seniors are crossing the ditch bridge on Kewanee and their car gets caught in the flow.
What once was 22 acres of low lying land to the east now has 30000 cubic yards of fill dirt and 134 very dense, lot-line to lot-line homes. This land, instead of absorbing flood waters, is now adding to the local floods. Dry Creek Ditch, instead of overflowing to the east, now has been channelized and contained. The water has no where to go except into the park, around the school and into people's homes.
Can this happen? Yes! Does this need to happen? No!.
As the City of Boulder states: Boulder is the #1 flash flood risk community in Colorado and, as the recent South Boulder Creek flood study attests to our neighborhoods are not immune from this risk. As Alan Taylor, city flood manager, has stated concerning a 500 year flood event:
Its on the order of comparing it to a Level 5 hurricane.Even the 100 year flood will have a serious impact on the neighborhood. According to the latest study results the Hogan/Pancost area is in the South Boulder Creek flood plain. Based on these models up to 1/2 of this parcel would be affected by a 100 year general storm.
The 500 year thunderstorm model is much worse. Here much of the western half of the property suffers from major flooding, Manhattan Middle School is cut off by flooding on all sides and the downstream neighborhoods (e.g., Tenino) are at major risk.
1969 FloodThis area can and will flood as attested to by the 1969 flood event. The following aerial photograph was taken from just east of the Hogan/Pancost property looking west. The other pictures were taken from the east end of Kewanee along the direction of the red lines.
Here is a picture taken west along south boulder road:
There was also a flood on this property in 1973.
According to residents who lived here at the time, these pictures were taken after the peak of the flooding. They describe the flood:
The run-off from the heavy rain on Sunday evening caused Dry Creek #2 to over-flow spilling out into at least an 175 foot wide river in the field East of us. Needless to say, Kewaydin Meadows was surrounded on all sides by water. Just East of Manhattan Drive on South Bldr. Road, the water flowing over So. Bldr. Rd. at one point reached a depth of about 2 feet. The area between Thunderbird Drive (now Foothills Parkway) and Manhattan Drive on Baseline road was closed for several hours due to the high flowing water. Probably in excess of 2 to 3 feet in a few places.In general, the entire area is essentially part of the South Boulder Creek floodplain. Even relatively small storms can cause extensive flooding as demonstrated by the image below. This shows Manhattan flooding after a 1/2 inch rainstorm in August 2007.
One or two basements on the North side of Kewanee Drive flooded with several feet of water. Water from the ditch backed up on Kewanee Drive to about the middle of Cimmaron Way.
Since the 1969 flood time there have been many changes to the area, some good:
- Viele Lake Drainage Channel was not complete by the time of the 1969 flood. This serves to route flood waters east towards South Boulder creek. However, as we discuss below the effectiveness of the channel depends in large part on the culvert under South Boulder road remaining open.
- A berm was built on the (now) CU South property. This berm is what triggered the current flood study and serves to greatly exacerbate flood impacts.
- Greenbelt Meadows was developed on fill thus removing 20 acres of flood storage.
- Much more development has occured along South Boulder road.
- The soccer fields were put in to the north, focusing flood waters onto the Pancost property and towards Manhattan Middle School.
Floodplain DevelopmentThe scope of the potential flood impacts is shown below. These images show a detailed view of the property with the 100 year and 500 year flood plain maps.
Let's take a closer look at the area. The following map shows the western edge of the property overlaid with the estimated 100 year flood flood plain. Note the flooded area is 100-150 feet wide and encompasses the entire property of 18 proposed homes. Each pair of pink triangles shows two elevation measurement points for each adjacent house in Keewaydin Meadows.
The values show the elevation difference between the pink triangle location near the home and the corresponding location on the development site. This shows the 2 northern homes are actually .5 to 1.9 feet below the corresponding location on the site. The other homes range from being 1.2 to 3.7 feet above their corresponding point on the site.
It should be noted that the site elevations we are using are the current land elevations, not the flood elevations. The flood will bring an additional height thus decreasing the elevation difference. The light green shows water up to 6 inches deep, darker green 0.5-1 foot, dark green 1-1.5 feet, yellow 1.5-2 feet.
The following pictures show water backing up on the streets to the west in the Spring 2005 from normal ditch operations. Now, imagine what this area will be like in a major flood.
The proposal dramatically decreases the flood storage capacity of the site. The 140 foot wide flood swath along the western side of the property holds, on average, 6 inches of flood water. Since it is 1400 feet long this area holds 90000 cubic feet of flood water (~2.5 acre feet). The central area of the site has an estimated 2 acres of flood coverage (or 1 acre foot at 6 inches depth).
The above ditch will result in flood flows of approximately 10 feet per second at 2 feet deep. The City's standards for a "high hazard" flood zone define one to be where the product of the depth times the speed is greater than 4. This drainageway will result in a flood zone where this product is 20, 5 times the standard.
Essentially they are creating a barrel of a gun pointing right at the school, park and downstream homes. The ditch, which through the development is engineered as a major flood drainageway, enters the park and school area and goes underground through the following culvert which would surely back up.
Boulder's Flood PoliciesThe city has a set of guiding principles as pertains to floodplain development. Clearly allowing development on this parcel violates the spirit of the city's own principles:
Local guiding principles for flood management include:If these principles, summarized by the city's own philosophy of "preserve floodplains", are relevant anywhere they are relevant to the Hogan/Pancost property. We believe that is entirely "possible" to preserve this floodplain. The land is in the county, the city does not have to choose to annex it. Area II designation in the BVCP is not a promise of annexation and development. Not annexing this property would in no way be considered a taking. However, annexing and allowing development could be considered a taking - a taking of the adjacent homeowners safety of their selves and their property.
1. Preserve Floodplains - where possible to recognize the beneficial functions of floodplains for hazard reduction, water quality enhancement, wetland protection, wildlife habitat, riparian corridors, recreation, alternate modes travel, environmental relief, aesthetics, greenway areas.
4. Prevent Adverse Impacts and Unwise Uses in the Floodplain - through appropriate regulation and land use, open land preservation and acquisition, multi-objective planning, relocation or elimination of high hazard structures, prohibiting unacceptable encroachments, and establishing ongoing maintenance practices that preserve and enhance environmental functions.
The city faces a large number of problems with existing development in floodplains. Why add to these problems by allowing new development of such scale?
Section 9-9-9 Floodplain Development Permits of the Boulder Revised Code states:
(c) When reviewing an application for a permit, the city manager shall determine which portion or portions of the floodplain are affected by the particular development request and shall then apply the provisions of Sections 9-9-6, "Regulations Governing the Floodplain," 9-9-7, "Regulations Governing the Conveyance Zone," and 9-9-8, "Regulations Governing the High Hazard Zone," B.R.C. 1981, as applicable. The manager also shall determine whether the application meets the intent of this chapter prescribed by Section 9-9-1, "Legislative Intent," B.R.C. 1981, after considering the following factors:
(2) The effects upon lands upstream, downstream, and in the immediate vicinity;
(3) The effects upon the one hundred year flood profile;
(4) The effects upon any tributaries to the main stream, drainage ditches, and any other drainage facilities or systems;
(7) The potential danger to persons upstream, downstream, and in the immediate vicinity;
(8) Whether any proposed changes in a watercourse will have an adverse environmental effect on the watercourse, including, without limitation, stream banks and streamside trees and vegetation;
(12) Whether safe access is available to the property in times of flood for ordinary and emergency vehicles;
(14) Whether the cumulative effect of the proposed development with other existing and anticipated uses will increase flood heights; and
(15) Whether the expected heights, velocities, duration, rate of rise, and sediment transport of the floodwaters expected at the site will adversely affect the development or surrounding property.
In light of these guiding principles and regulatory processes some of the questions that need to be asked are:
- Will this development exacerbate the flood impacts on the surrounding
neighborhoods and school?
Manhattan Middle School and many homes are immediately downstream of the
parcel and face even greater risks of flooding.
While the developer may be able to mitigate
the impacts on this development will it make it worse for others.
- Will development change the results of the ongoing South Boulder Creek flood study?
How will raising the level of the land and adding a large number of houses change
the hydrology and hydraulics of this area. Many homes in southeast Boulder
are already facing serious flooding issues, lets not make it worse.
- Will development here preclude potential mitigation efforts?
As Bob Cole (head of the Planning Department's Land Use Review)
stated 2 years ago:
It is possible that some portion of the site may be needed for additional flood storage capacity in the future, but that won't be known until completion of the flood study and mitigation plan. To annex the site and approve a development plan in advance of the flood study completion may foreclose important flood mitigation options for the future.See here for further discussion of the mitigation issue.
South Boulder Creek Flood StudyWe believe there are some serious issues with the use of the flood South Boulder Creek flood study results to characterize flooding in the Hogan/Pancost area. If one compares the flood plain maps produced by the flood study with actual ground features there are some serious discrepencies base on:
- Model DEM resolution.
- A fixed blockage model.
- Not taking into account land micro-features.
- Lack of basic what if scenario planning.
Model DEM ResolutionThe flood study is based on a 1 dimensional model that represents stream channels, ditches and other water ways. This model of the water flow then provides the basis for modeling the water on a 2 dimensional topography of the land. This 2D model is based on a one foot Digital Elevation Model (DEM) that the city created in 2003 for this flood study. This 1 foot DEM is a grid with 1 foot spacing. However, the actual flood modeling was done with a 4 Meter DEM that was based on this 1 foot DEM. So, it is gauging flows with an elevation point every 4 meters. This is done because of the computational complexity (i.e., time) of running the model with a higher DEM. At the large scale of the many square miles of the entire South Boulder Creek basin this 4 Meter DEM is probably of high enough resolution. However, we have doubts as to its accuracy when applied to a smaller scale area such as the 24 acres of the Hogan/Pancost property. Many important topographical features such as drainages, ridges, etc., can be lost within 4 meters.
Micro-featuresOne of the fundamental flaws of the flood study as it pertains to Hogan/Pancost is that the model does not represent the myriad number of "micro-features" such as ditches in the area.
As the above image shows the Hogan/Pancost property is surrounded on all sides and bisected by irrigation ditches. However, the ditches shown in light blue are not used in the flood study model. While some of these are relatively small lateral ditches (though they do contribute to flooding) others are fairly large and can and do transport a large amount of water. The transportation mechanism of flood waters that these ditches provide is not captured by the model, thus the actual extent of the flooding produced by the model is greatly reduced.
For example, the ditch along the northern border of the property comes directly from South Boulder creek. As can be seen in the map image above, it leaves the creek at a fairly sharp bend where the alignment of the ditch follows the alignment of the creek. This ditch is almost as large as the Dry Creek Ditch #2 and offers a more ready ingress to flood waters from its source at South Boulder creek. Why isn't this ditch captured in the model? As can be seen in the image below it carries a substantial amount of irrigation water in a normal year. However, as noted in the flood map above the area centered around 55th street is shown as being dry. This area is 200 meters down from South Boulder Creek. Furthermore, the headgate is positioned at a critical position along the creek where the creek veers to the east and the alignment of the ditch continues to the northwest. Now assuming that a 500 year thunderstorm flood event is of fairly large proportions how can this area possibly not be under water?
Another example of the suspicious results of the flood study is shown below. On the left is a local map of the 500 year thunderstorm showing 55th street, north of South Boulder road. Note that there is no water along 55th street and a very small amount of flooding in the pasture. The dark blue circle near the house highlighted in light blue is the site of the picture shown on the right. This picture was taken on May 24th during a small scale flood event due to the irrigation ditch that flows along 55th street. So, a very small scale event produces flooding in an area where the model of the 500 year thunderstorm shows as being dry. Do we trust the model or do we trust our own eyes?
And finally, the picture below shows the results of recent flood irrigation of the property. There is 6-10 inches of standing water across approximately 1+ acre of the site. However, this area, according to the 500 year thunderstorm flood map (a very large flood event), is dry! We readily admit that we are not flood engineers but we do know one thing: water flows downhill. There are no major flood control structures along the set of ditches. If water can flood the central part of the property from simple irrigation why does it not also flood during a 500 year thunderstorm event?
The videos below were taken in Summer 2006 of this same area. This is not a trivial amount of surface water:
Local FloodingThe flood study also does not adequately take into account local storm based flooding and minor tributary flooding, i.e., flooding that occurs due to precipitation on the immediate area. FEMA does not require taking into account the impacts of local precipitation when developing the Flood Insurance Rate Maps (FIRM). However, if the city is truly interested in understanding and dealing with flood threats then local storms need to be taken into account. The impact of local precipitation can and will have a serious effects both on its own and in conjunction with larger basin wide flooding. Many of us have witnessed the impact of a local storm. Ditches flood, gutters run over, water flows down stream, the ground becomes saturated. How will this affect the flooding behavior from upstream? These questions are especially releveant for the general storm event where there is widespread precipitation.
BlockageBlockage in the model presents a signifcant problem. So far, the flood study assumes 100% blockage of culverts less than 36 inches in diameter but, as yet, does not take into account blockage of larger culverts and bridges. As noted here it is planned to incorporate these into the model:
Yes, the study considers blockage at culverts and bridges, and satisfies FEMA guidlines. The hydraulic model assumes that all culverts along main and secondary channels with a diameter less than 36 inches are blocked entirely. Maintenance of these structures is variable, and even minor events produce debris that significantly reduces culvert performance.The study materials state (January 2004):
Larger structures also assume blockage by debris or sediment accumulations based on an assessment of expected impacts given size and location. For example, we may assume that medium to large structures will have a 30% degree of blockage. Larger mutiple span structures that are more prone or less prone to debris accumulation are to be modeled using a more specific assessment based on their span size and peer placement relationship.
Blockage has been brought up but there is not a perfect way to study this issue.We agree with this assessment. For example, the current model assumes all small (<36in) culverts are completely blocked. The problem with this assumption is that it may serve to dramatically change where the model predicts 1-dimensional flows to take place.
An on the ground inspection easily shows the importance of taking into account culvert and bridge blockage. We believe that the flood delineation may dramatically change when the full blockage potential is taken into account. This is especially true for the Viele Lake Drainage Channel. This is a fundamental component of drainage system and serves to route flood waters away from the Manhattan/Dry Creek Ditch/55th Street corridors. However, if it were to block at the underpass on South Boulder Road there would be a large amount of spill over that flows through the neighborhoods to the north.
We also believe that due to its complexity and importance, the drainage channels and ditch networks along the Viele Lake Drainage Channel deserve special attention. The current 100 year general storm maps show 5-7 feet of water in the Viele Lake channel and whether or not there is blockage there is great risk of serious spillover to the north. The complexity of this area may not be captured well by the models. For example, the ditch that flows along the north side of South Boulder Road shows little or no flooding.
"What if" scenario planningAs the flood study states:
Where the flooding will occur and the depth of the flooding at a particular location will vary depending on a wide variety of factors.Most of us, when engaged in a decision making process where there are unknowns and a range of potential outcomes and risk try to do best-case/worst-case scenario planning. This helps frame the uncertainty and gives us a tool to make better decisions. We do not believe that the SBC flood study has taken this approach. The only example of this (albeit an important one) is modeling with and without the berm on the CU South property. However, the complexity and cost of doing a single model run seems to have precluded doing best-case/worst-case modeling. It seems as though that the modeling team, after thoughtful deliberation, just came up with some parameters:
- Land roughness factors.
- What ditches to include in the 1D model.
- Blockages of culverts.
- Grid size, homogeneous grid.
The flood study web site goes on to state:
We will make sure the study is not adopted if it is not correct.Our question is is how does one measure correctness? Is it correct because they think they have the climatology right and the MIKE model produced some results that seem reasonable in its own context?
Flood HistoryFrom a memo to the Water Resource Advisory Board we see the following:
3. The most well-known South Boulder Creek floods of record (September 1938 and May 1969) were more significant than perhaps realized. The 1938 flood involved a rare hurricane driven thunderstorm that produced rainfall amounts seen an average of once every 500 years. The 1969 flood involved a 5-day general storm that produced rainfall exceeding 100-year levels, much of which was fortunately captured in Gross Reservoir. Experiencing two significant storms in 65 years may seem unusual but we are finding that such occurrences are not unprecedented.The above highlighted sentence strikes us as odd. By saying that signifcant and frequent storms are not unprecedented they are in essence saying that the the probability of occurence of these size storms is much higher than has been previously estimated. Which, says that the estimations for what is a 100 year or 500 year storm and flood event is in error.
On the one hand, we have direct historic evidence of major flooding a number of times over the past century. These storms were real - and they surpassed the estimates of what is a 100 year or 500 year flood. Again, what should we believe - our own eyes and the historic record or the results of a statistical analysis.