Long-term trends Average annual temperature and precipitation levels observed in this area have been increasing.
Small storms matter Around 90% of all rainfall events in Central Iowa are equal to or less than 1.25 inches. Since most rainfall is generated by these smaller events, most pollutant loads are washed off surfaces and carried to streams by these storms.
Moderate storms matter too In this watershed, storm events of greater than 2.67” in a 24-hour period occur about once a year. These storms which occur about once a year on average are larger than 98% of local rainfall events. These storms often cause local flash flooding and can destabilize streams due to erosion caused by the rapid “bounce” in stream flows following such storms.
Runoff is increasing Average annual streamflow in Walnut Creek has increased 37% since 1981.
Seasonal flow patterns Highest flow rates have been most commonly observed from late April through June. Some very significant flood events have also occurred in July and August.
Land use impacts stream flow Streamflow has increased more intensely than precipitation, indicating that land use changes are influencing streamflow increases.
HOW DO THESE CONCEPTS INFLUENCE DEVELOPMENT OF THE PLAN?
Climate factors such as temperature and precipitation ultimately impact pollutant loads. Certain pollutants are more likely to be present after storm events. Some are more common when temperatures are higher. Higher rainfall amounts could lead to increased flood risks. Such risks also are influenced by changes in land use.
Records were used to review cycles and trends in average annual temperature. A normal annual mean temperature was calculated using the average annual temperature readings over the previous 30 years for any given date through the period of record (1893-present).
Over the past 50 years, January has generally been the coldest month of the year, with a mean temperature of 22.90˚F. July has typically been the warmest month of the year. The highest temperature recorded at the Des Moines Airport was 110˚F on July 25, 1936 and August 4, 1918. The minimum temperature reading of -29˚F was registered on January 12, 1912.
An analysis calculated a normal growing season length using the average length of season over the previous 30 years. Even though normal annual temperatures increased until 1942, the normal growing season length generally decreased until 1972. Since that date, the average growing season has generally been on the rise, although a downtrend did occur between 1982 and 1992. The normal growing season length in 2014 was calculated to be 177.5 days (based on period form 1985-2014).
Precipitation data was reviewed in a manner similar to temperature information. Annual precipitation totals observed at the Des Moines Airport over the period of record range from 17.08” (1956) to 55.89” (1993), with an average of 32.14” over the entire period of record. A normal annual precipitation depth was computed using the annual precipitation over the most recent 30 years through the period of record. Up until 1982, the annual normal precipitation generally remained within 0.5” of the average of the overall period of record. But since that time, annual rainfall has been trending upward. Normal precipitation values peaked in 2011 at 36.15.” The normal value dropped slightly through 2014 to 35.23,” as the three consecutive wet years of 1982-1984 dropped out of the computation of the 30-year normal. Even so, the normal annual precipitation remains more than three inches above the longer-term average. There appears to be a clear local trend of increased precipitation over the past three decades. Six of the top eight wettest years on record have occurred since 1982, while none of the driest years on record have occurred during that same period.
Normal precipitation levels are lowest in January, rising through the spring to highest levels during the month of June. Precipitation levels remain elevated through July and August, falling off significantly through the fall months Engineers use published studies on local rainfall rates in the design of storm sewers, culverts and storm water management practices. In the past, local design standards have used information from Technical Paper 40 (1) (1961) and Bulletin 71 (2) (1992) rainfall depths and rates for short-term events (time periods ranging from a few minutes to up to 10 days). Recently, NOAA issued their Atlas 14 (3) rainfall data set, which included more recent precipitation data to establish these rates. SUDAS (Statewide Urban Design Standards and Specifications) and ISWMM (Iowa Stormwater Management Manual) have adopted these values for use in design. Values listed in Atlas 14 are generally similar to or higher than those listed in previous studies, meaning that storm water facilities now are expected to be designed to handle runoff generated by more rainfall than expected in the past.
Available Streamflow Gage Data
Stream flow data has been collected at a USGS gaging station located near the 63rd Street Bridge on the border between Des Moines and West Des Moines (USGS 05484800). Data collection began in October of 1971 and continues through the present day. At this location, Walnut Creek is collecting runoff from an area of 78.3 square miles (95% of its entire watershed).
Stream flow varies greatly from year to year. Since 1972, annual flow volumes have ranged from 334 million cubic feet in 1989 to 5.64 billion cubic feet in 2010. To put that in perspective, the annual volume of flow from 2010 would be enough to fill 21,000 large water towers (assuming each tower could hold 2 million gallons). A general upward trend can be observed in runoff volume. The value of annual flow based on an average over the previous 10-years has increased from 1.9 billion cubic feet in 1981 to a high value of 2.6 billion cubic feet in 2014, an increase of 37%.
Daily average flow rates in Walnut Creek have ranged from very little flow to 5,100 cubic feet per second on August 9, 2010. The total flow on that single day of 441 million cubic feet would have exceeded the flow for the entire year of 1989, and would have been equal in volume to more than 1,600 large water towers. However, such large flows are uncommon.
Average daily flow rates have exceeded 1,000 cubic feet per second for only 67 days over a period of more than 43 years (less than 0.5% of all days). The average daily flow rate over the entire period of record is 26 cubic feet per second, or a daily volume of 2.3 million cubic feet.
A typical flow curve for Walnut Creek has been constructed by reviewing average daily flow data collected from October 1, 1971 to August 5, 2015. Over this period, highest flows are most often observed between late April and the end of June. Flows typically decline through July, although higher flows have sometimes occurred during periods in early and late August. The typical flow curve is constructed by using a 30-day average (looking 15 days before and after a given date). This has the effect of smoothing out the high and low values into a more regular pattern over the course of an “average” year.
Conversion of Precipitation to Streamflow
The amount of precipitation which is converted to streamflow has varied greatly. Comparing precipitation observed at the Des Moines Airport and data from the stream gage at Walnut Creek, we can find that in 1989, annual stream flow volume was only 7% of the volume of that year’s precipitation. Contrast that with 2010, when stream flow equaled nearly 60% of annual precipitation.
Many factors appear to influence this value which are hard to separate at this level of study:
-Was that year more wet or dry? Streamflow percentages are higher in wet years than dry years.
-Was the previous year wet or dry? A previous dry year may have lowered groundwater and surface moisture levels, resulting in less runoff.
-Did precipitation come in large storm events or more steady over time? A higher percentage of runoff is expected during heavier rainfall.
-How have land uses changed? Development activities install impervious surfaces (hard/solid surfaces that won’t slow or hold water like rooftops and driveways) and development compacts remaining open spaces through mass grading. These factors lead to higher levels of runoff.
With the information that is available, it is difficult to draw conclusions on how each of the items above influences the percentage of precipitation that is converted into streamflow. From the information on hand, we can state the following:
- The portion of streamflow that is converted to runoff increases with annual rainfall volume and increases in impervious land cover.
- Reviewing averages over the previous 10-years from a given date, the proportion of precipitation converted to runoff has increased from an average of 30.4% in 1981 to 37.6% in 2014.
- Normal annual precipitation has risen from 30.61 inches in 1981 to 35.23 inches in 2014 (based on the average precipitation of the previous 30 years from the given date).
- Urban land cover has increased greatly, from 35% to 43% of the total watershed area (a change of 4,300 acres) between 2001 and 2011.
Flood Risk Potential
Flooding is a key concern within this watershed, frequently discussed at board meetings and open houses. A flood event that occurred during the planning process intensified the focus on this issue. Flooding occurred during the morning hours of June 25, 2015. This event occurred when more than five inches of rain fell over portions of the watershed beginning late in the evening of June 24. Most of the rain fell over a three to five hour period, depending on location. Urban development has occurred within many flood prone areas. In some cases, structures have been located in areas where they are frequently flooded including residential, commercial, industrial and public buildings. Flood risks in this area have been evaluated multiple times. Studies of flooding are developed into maps that demonstrate different levels of expected risk along major stream corridors (Flood Insurance Rate Maps issued by FEMA). These maps are intended to identify the need for flood insurance to be purchased by property owners.
These maps typically identify key features of a flood plain:
Base Flood Elevations—Detailed studies may identify expected high water elevations caused by a 1% annual exceedance probability flood. This is a flood event, expected to have a 1% chance of occurring in any given year, or over very long periods of time would be expected to happen once every 100 years on average. This has commonly been referred to as a “100-year flood event,” although the phrase “1% annual exceedance probability” is now the preferred terminology. Aerial photos from flood event—June 25, 2015.
Floodway—A zone where grading or structure construction is most limited. This zone is intended to be kept clear of obstructions. In theory, the flood plain could be completely filled on either side of this zone and the expected result would be a one foot rise in base flood elevations. Not all maps identify the floodway for a given stream, but they are usually identified on larger streams or in areas where more detailed studies have been completed.
Flood Fringe—Maps often identify areas outside of the floodway which show areas expected to be covered by the 1% or 0.2% annual exceedance probability flood events. The 0.2% probability event has commonly been called the “500- year flood event.” In some areas, these maps are associated with studies that provide detailed cross-sectional information of the flood plain. Such studies may include expected flood elevation profiles for other more commonly occurring storm events (i.e. 2%, 10% exceedance probability, etc.). It is important to understand that localized flash flooding can occur outside of areas with mapped flood risk. This can be caused by clogged inlets or storm sewers and culverts; overloaded storm sewer systems, blocked overflow paths and urban small stream flooding.
At the USGS gaging station at the 63rd Street Bridge, minor flooding impacts are expected at a gage height of 14 feet. At this location, major impacts due to flooding are expected when water exceeds a gage height of 17 feet. Gage height data has been collected every year since 1972. Over that 44 year period, there have been 14 years (31.8%) where a gage height above 14 was recorded. Water levels have exceeded a gage height of 17 feet in eight (18.1%) of the years on record. This indicates that a flooding to a gage height of 17 feet may be expected once every five years on average.
Hydraulic Modeling and Flood Inundation Mapping
To begin the hydraulic and floodplain mapping investigation for the Walnut Creek Watershed Management Plan, data and models from three previous studies were reviewed. These previous studies included hydraulic and hydrologic analyses for the North Walnut Creek watershed, hydraulic and hydrologic analyses for the Walnut Creek watershed, and updated FEMA floodplain mapping for North Walnut Creek and Walnut Creek. These studies and models were produced using rainfall data published in 1992 (Bulletin 71), which was the most recent rainfall data available at the time these studies were completed. These studies were eventually incorporated into a FEMA floodplain mapping update that was published as a preliminary study in June 2015. After a public comment period, they are scheduled to become effective in the winter of 2016.
This study further updates hydrologic and hydraulic models as well as inundation mapping for Walnut Creek and North Walnut Creek using more recent rainfall estimates. This plan uses NOAA Atlas 14 data. The Atlas 14 rainfall estimates take into account a longer period of statistical data than Bulletin 71 and include rainfall data through the 2013 water year. In general, in Iowa, the rainfall estimates for NOAA Atlas 14 have increased when compared to Bulletin 71. This is also the case for the Walnut Creek and North Walnut Creek watersheds. This increase in rainfall estimates, in turn, increased the peak flow estimates, water surface elevations, and floodplain extents.
The results of the analyses using the Bulletin 71 estimates and the NOAA Atlas 14 estimates were compared. Points of interest are outlined for each flooding source below.
Walnut Creek 1% Annual Exceedance Probability:
- 73 acres greater floodplain area for the Atlas 14 based mapping over 18 miles of stream length
- An average of 34 feet wider than the preliminary mapping or 17 feet on either side of the stream
- Smaller in some areas due to some of the additional area being within backwater fingers
North Walnut Creek 1% Annual Exceedance Probability:
- 16 acres greater floodplain area for the Atlas 14 based mapping
- An average of 24 feet wider than the preliminary mapping or 12 feet on either side of the stream
- Smaller in some areas due to some of the additional area being within backwater fingers
The differences highlighted above are a direct result of increased rainfall estimates with the updated statistical data. However, as land use and management changes within the watershed, these increases in peak flow estimates could worsen over time. Since future meteorological changes are unknown, to mitigate risk and future losses, watershed and floodplain management becomes paramount.