Watershed Modeling
Images from the USGS Water-Resources Investigation Report 2010-5007, which describes the analyses conducted by USGS for the Ipswich Targeted Watershed Grant project

Location: Watershed-wide analysis

Purpose: Evaluate the potential for LID and innovative water conservation techniques to improve base flow within the Ipswich River, if the techniques were to be applied on a wider, regional basis.

Description: In 2000, the U.S. Geological Survey (USGS) used the Hydrologic Simulation Program--Fortran (HSPF) to develop a watershed-scale precipitation-runoff model of the Ipswich River. The model simulates the watershed’s hydrology and water-use patterns to relate pumping, precipitation, and land use to streamflow (Zarriello and Ries 2000, p. 1: click here for the full USGS report).  In addition to simulating existing conditions, the model may be adjusted to simulate theoretical alternative water pumping and land-use patterns.

 

To develop a “baseline” to represent current conditions, USGS updated the model published in 2000 to reflect substantial recent changes to pumping regimes in two headwater communities.  Specifically, the updated baseline reflects that Reading recently eliminated its Ipswich Watershed withdrawals and began purchasing its full water supply from the Massachusetts Water Resources Authority (MWRA), which derives its water from reservoirs in central Massachusetts.  Similarly, Wilmington is reducing its Ipswich Watershed withdrawals by approximately 1 million gallons per day (MGD) during the summer and replacing this volume with water purchased from the MWRA.  All other aspects of the original model remain unchanged for the updated baseline.

A Technical Advisory Committee (TAC) consisting of hydrologists, water suppliers, the Ipswich River Watershed Association, and state and federal agency representatives was convened to guide the development of theoretical scenarios to be evaluated.  Streamflow impacts from the following scenarios were modeled:  

  • Full Build-out – This scenario simulated the conversion of all developable land in the watershed to its zoned use.
  • LID Retrofit – This scenario simulated wide-scale implementation of LID retrofit techniques in the upper watershed, the most densely developed part of the watershed.  Specifically, the effective impervious area associated with all developed land use was reduced by 50%, and recharge was increased accordingly.
  • Water Conservation Scenario – To simulated basin-wide application of the water conservation pilot strategies , water savings calculated for the pilot projects were scaled up to the town-wide level, using appropriate data for each town.  Tables pdf format of    Tables   Withdrawals at public water supply sources were then reduced to reflect these town-wide savings.  Depending on individual town characteristics and overall water use patterns, these per-town reductions ranged from 0.27 million gallons per month to 4.4 million gallons per month and from 1.4% to 8.5% of a town’s total monthly withdrawal.
  • Local-scale simulations – A series of simulations for a hypothetical 100-acre parcel was conducted, to evaluate more localized streamflow impacts resulting from land-development and LID.  These simulations evaluated the impact of: 1) natural geology; 2) preservation of open space; and 3) reduction of effective imperviousness across various land uses, on streamflow immediately adjacent to the 100-acre parcel,

Data Collection and Analysis: For detailed information on the HSPF model and the modifications made to simulate the scenarios described above, please see the USGS Scientific Investigation Report 2010-5007

 

Key Results and Conclusions:

  • Updated Baseline - Results from the updated model baseline suggest that the elimination of Ipswich Watershed withdrawals by Reading and the reduction by 1 million gallons per day of withdrawals by Wilmington notably improves low-flow conditions in the Ipswich River, most significantly in the headwater reaches adjacent to the Reading and Wilmington wells.
  • Build-Out – The build-out simulation resulted in negligible additional impacts to streamflow for most of the tributaries and stream reaches in the watershed.  The largest impact was observed in a stream that drained substantial amount of undeveloped land that is zoned for industrial and commercial land uses.  For this stream, the build-out condition resulted in higher high flows (more flooding, scouring, and erosion) and lower low flows compared to current conditions.  The reason that so little impact was observed from the build-out scenario for most of the watershed is due to the fact that most of the developable land in the watershed is zoned for low-density housing, which tends to have little effective impervious area. Therefore, the build-out condition does not represent a substantial change in effective impervious area over the whole watershed, relative to today’s conditions.  

    It’s important to note, however, that while the model indicates that streamflows in major tributaries and the mainstem are unlikely to be directly impacted from new development, streamflows are very likely to be indirectly affected, if the water demand from the new development is met with Ipswich Watershed sources, which will increase pumping beyond the current rates.  Additionally, the model was not able to account for decreases in water quality that may results from build-out conditions.
  • LID Retrofit – The LID retrofit simulation similarly resulted in negligible improvements to streamflow conditions in the major tributaries and stream reaches.  The stream reach most dramatically affected by simulated LID retrofits drains a portion of the upper watershed that has a high percentage of effective impervious area (22%), due to the presence of lots of commercial and industrial parcels.  For this stream, a 50% reduction of impervious area in the vicinity led to decreased high flows (less flooding, scouring, and erosion) and higher low flows (the retrofits eliminated periods of time when the riverbed would go dry).  Most of the other tributary and stream reaches drain areas with effective impervious areas of 9% or less.  These levels of effective impervious area are low enough that a reduction by half does not amount to a large absolute change in watershed flow conditions.
  • Water Conservation Scenario – The simulated reductions in town-wide pumping, ranging from 1.4% to 8.5%, reflective of the hypothetical wide-spread adoption of the pilot water conservation strategies had very modest effects on flows in the tributaries and mainstem reaches.  By comparison, a simulated reduction of 20% at all public water supply withdrawal sources in the watershed resulted in a slightly more measurable mitigation of extremely low flows.  In sharper contrast, the simulated ceasing of all public water withdrawals in the watershed resulted in dramatic improvements in streamflow in many tributaries and mainstem reaches.
  • Local-scale land use simulations – The local scale modeling demonstrated that small streams in the vicinity of parcels that undergo significant development undergo substantial impacts to flow conditions.  The conversion of the 100-acre parcel from forest to commercial land use dramatically increased flood flows and caused severe low flows.  Conversion of the forest to low-density or high-density housing increased flood flows somewhat, but tended to raise the volume in the stream during low-flow periods, too.  This can be attributed to the clearing of trees, which soak up groundwater during the growing season.  Since residential landscapes draw less water from the subsurface groundwater stores than forests do, groundwater contributions to streamflow can actually increase when forests are converted to residential development. 

    The local scale simulation series also investigated the impact of clustering development on a portion of the lot and preserving the natural forest on the area that remained.  In most cases, clustering succeeded in significantly reducing the impacts to streams caused by more uniform development.  Lastly, the local scale simulation explored the influence of natural geology on the above trends, and the results suggest that land use changes that occur over subsurface conditions dominated by sand and gravel tend to cause more significant disruptions to natural streamflow than land use changes over subsurface conditions dominated by glacial till.  As a result, introducing LID to mitigate the streamflow impacts of development makes a more substantial difference in areas dominated by sand and gravel.

 

Final Reports: 

Watershed Modeling: U.S. Geological Survey