Appendix A

Appendix A - Technical Appendix

Technical Memorandum

Methodology for HHPLS P-Load Model Application to

Development of Subwatershed Rules under Performance-Based Management

Details how the MCWD 2003 Hydrologic, Hydraulic, and Pollutant Loading Study (HHPLS) modeling was used to develop the phosphorus load and annual water volume assumptions for this plan.  It also explains how Ultimate Development conditions were calculated.

Ultimate Land Use Models

Spreadsheets for the major receiving waters in the watershed, showing the 2000, 2020, and Ultimate Development modeled conditions and resulting lake response.  These models were used to calculate the phosphorus load reductions necessary to attain in-lake total phosphorus concentration goals.

Ultimate Volume Control Model 

Spreadsheets summarizing annual runoff data the HHPLS and Ultimate Land Use modeling.  The spreadsheets show the calculation of the estimated annual volume of runoff that could be expected under various infiltration scenarios – infiltrating the first 0.5”, 0.75”, 1.0” of runoff.  The purpose of this modeling was to determine how much of the new runoff volume that might be generated from new development could be captured on site through small-event infiltration, and how much new volume would run off.

Technical Memorandum

MCWD Management Plan Revision

Calculation of Subwatershed Phosphorus Load Reductions

 

Explains how LGU expected phosphorus load reductions were calculated.

 

Technical Memorandum

MCWD Management Plan Revision

Subwatershed Phosphorus Load Reductions Compared to HHPLS

 

Explains how the LGU expected Phosphorus Load Reductions in this Plan compare to those considered in the HHPLS 

 


 

 

The 2003 MCWD Hydrologic, Hydraulic, and Pollutant Loading Study (HHPLS), through an extensive public input process, established lake water quality goals (later adopted by the Board) and identified strategies for reducing phosphorus in modeled lakes within the MCWD.  For many lakes one of the strategies was a phosphorus load reduction that could result from generalized application of BMPs in residential and other developed land uses across a subwatershed.  The expectation was that local governments could achieve these reductions through the application of various BMPs as opportunities arise. 

 

The HHPLS included these types of general reductions by the LGUs for some subwatersheds, but not for all.  For consistency and fairness, the District has proposed in each subwatershed plan of the 2006 MCWD Water Resources Management Plan a required reduction by the LGUs in the subwatershed of the phosphorus load contributed by existing land uses.  The requirement is a 15 percent reduction in loading from existing residential land use; 25 percent from agricultural land use; and 10 percent from other developed land use.  The exception to this requirement is where a TMDL has determined what the lakeshed load reduction should be; the TMDL reduction is required instead.

 

This reduction can be accomplished through:

  • · Application of BMPs such as additional street sweeping, local water quality ponds, reduction of runoff volume, rain gardens and infiltration swales, and agricultural BMPs that reduce erosion or treat runoff or drain tile discharge;
  • · Prevention of future load increases through the conservation of lands previously identified for development; or
  • · Achieving load removals in excess of the minimum required.

 

The LGUs must identify in local water management plans specific steps to accomplish these minimum reductions.  The LGUs must also annually report to the District their progress toward accomplishing this requirement.

Calculation Method

 

This reduction was calculated using HHPLS modeling data.  That analysis assigned a land cover/use category to each subwatershed unit polygon using Minnesota Land Cover Classification System (MLCCS) data.  Expected volume and phosphorus loading could then be calculated based on land cover/use.  The subwatershed phosphorus load was then summed by land cover/use category for each subwatershed, or lakeshed where more than one lake is present in the subwatershed, and percentages applied to calculate the proposed load reduction.  Where a subwatershed unit is landlocked, no reduction was calculated.

 

Because MLCCS data is not available for most of the Minnehaha Creek subwatershed, load reductions for that subwatershed had to be calculated in a slightly different manner.  Developed land use accounts for about 75 percent of the land area of the subwatershed.  To reflect the higher concentrations that typically run off impervious surfaces, 80 percent of the modeled load was used as the “base” from which to calculate the load reduction.   A 10% reduction closely meets the subwatershed load reduction to achieve the phosphorus concentration in Minnehaha Creek necessary to achieve the Lake Hiawatha TMDL reduction.

 

The calculated load reductions for each subwatershed or lakeshed are then partitioned between the LGUs in the subwatershed based on area within each drainage area.  The following tables illustrate the calculation for the Gleason Lake subwatershed.

 

Gleason Lake BMP Load Reduction Allocations Calculation

 

Rcvg

Other-Developed

Residential

Other-Vac/Ag

 Agricultural

 

Shed

Water

 P-load

10%

 P-load

15%

 P-load

10%

 P-load

25%

Total

GLC-1

Gleason

     3.673

0

  105.906

16

     2.776

0

          -  

0

16

GLC-2

Gleason

     0.506

0

    26.954

4

     4.495

0

     0.002

0

4

GLC-3

Gleason

     0.596

0

  122.824

18

     0.617

0

          -  

0

18

GLC-4

Gleason

    25.547

3

    35.409

5

    27.182

3

     0.231

0

11

GLC-5

Gleason

     1.368

0

  123.276

18

    12.782

1

          -  

0

19

GLC-6

Gleason

    14.216

1

    67.025

10

     1.239

0

     0.076

0

11

GLC-7

Gleason

     5.664

1

  109.953

16

     4.136

0

     0.038

0

17

GLC-8

Gleason

    11.646

1

    21.539

3

     1.542

0

     0.120

0

4

GLC-9

Gleason

    23.429

2

    82.752

12

     8.001

1

     0.078

0

15

TOTAL

Gleason

 

8

 

102

 

5

 

0

115

HL-1

Hadley

    16.526

2

  114.998

17

     0.859

0

     0.011

0

19

HL-2

Hadley

          -  

0

    37.027

6

     5.946

1

     0.122

0

7

TOTAL

Hadley

 

2

 

23

 

1

 

0

26

GLC-10

Creek

     0.625

0

    17.390

3

     0.438

0

     0.086

0

3

GLC-11

Creek

    16.990

2

    25.353

4

     0.390

0

     0.039

0

6

HL-3

Creek

     0.080

0

    11.599

2

     0.011

0

     0.056

0

2

HL-4

Creek

     0.050

0

    56.733

9

     0.803

0

          -  

0

9

HL-5

Creek

    20.283

2

    58.068

9

     5.236

1

     0.054

0

12

TOTAL

Creek

 

4

 

27

 

1

 

0

32

TOTAL

TOTAL

 

14

 

152

 

7

 

0

173

 

 

Rcvg

 

 

 

 

 

Shed

Water

Plymouth

Wayzata

Minnetonka

Orono

Total

GLC-1

Gleason

16

0

0

0

16

GLC-2

Gleason

4

0

0

0

4

GLC-3

Gleason

18

0

0

0

18

GLC-4

Gleason

11

0

0

0

11

GLC-5

Gleason

19

0

0

0

19

GLC-6

Gleason

11

0

0

0

11

GLC-7

Gleason

17

0

0

0

17

GLC-8

Gleason

4

0

0

0

4

GLC-9

Gleason

10

5

0

0

15

TOTAL

Gleason

110

5

0

0

115

HL-1

Hadley

19

0

0

0

19

HL-2

Hadley

7

0

0

0

7

TOTAL

Hadley

26

0

0

0

26

GLC-10

Creek

1

2

0

0

3

GLC-11

Creek

0

6

0

0

6

HL-3

Creek

0

0

0

2

2

HL-4

Creek

6

3

0

0

9

HL-5

Creek

3

8

0

1

12

TOTAL

Creek

10

19

0

3

32

TOTAL

TOTAL

146

24

0

3

173

 


 

 

The 2003 HHPLS included load reductions that could result from generalized application of BMPs in residential and other developed land uses as part of developing phosphorus load reduction plans for modeled lakes within the MCWD.  These reductions were calculated for some subwatersheds but not all.  District staff has asked Wenck to calculate reductions for those subwatersheds where no BMP allocation was made, on the same basis as those made in the HHPLS so that this requirement can be applied uniformly across the District.

 

We have reviewed the HHPLS and the P-load spreadsheets used by EOR in the development of that study.  We have found that those BMP load reduction allocations in the HHPLS were not calculated in the same manner for all subwatersheds.  

 

To provide consistency in those allocations, we have used EOR’s data and applied a uniform method to calculate potential reductions.  This method would calculate the load reduction as follows:

 

  • 15% of existing modeled P-load from residential land use
  • 10% of existing modeled P-load from other developed land uses
  • 10% of existing modeled P-load from land uses characterized as “vacant/agricultural” but where the MLCCS code indicates there is some amount of impervious cover indicating it is developed
  • 25% of existing modeled P-load from agricultural land uses

 

Because MLCCS data is not available for most of the Minnehaha Creek subwatershed, we have used the following method.  Developed land use accounts for about 75 percent of the land area of the subwatershed.  To reflect the higher concentrations that typically run off impervious surfaces, we used 80 percent of the modeled load as the “base” from which to calculate the load reduction.   We have calculated load reductions as both a 10% reduction and as a 5% reduction.  A 10% reduction appears to most closely meet the needs outlined in the Lakes TMDL for Creek concentration reduction.

 

Table 1 summarizes the basis it appears was used in the HHPLS to calculate an LGU load reduction, the reduction in the HHPLS, and the reduction as recalculated using the consistent method described above. 
Table 1.  Summary of BMP Load Reduction Allocations in HHPLS and as Proposed.

Subwatershed

Receiving Water

Basis for HHPLS Load Reduction Allocations

HHPLS Total BMP Load Reduction Allocation

Revised Total BMP Load Reduction Allocation

Dutch

 

15% of total P-load on all land uses, plus 10% of DL-3 for “shoreline BMPs”

44

28

Langdon

 

None except 10% of that part of LL-5 not tributary to the proposed pond area on W side of lake (50% if 55ug/L goal is selected)

10

20

Schutz

 

Same as proposed

14

17

Virginia

Minnewashta

Same as proposed, except included landlocked basins

30

29

Virginia

Same as proposed

24

TMDL

Christmas

 

None

-

10

Gleason

Gleason Lake

None

-

115

Creek

 

-

58

Long

Long Lake

None

-

118

Tanager Lake

None

-

38

Painter

 

None

-

157

Six Mile

Auburn E

15% of residential and 25% of agricultural

28

34

Auburn W

None

-

3

Lundsten

None

-

23

Marsh

None

-

31

Mud

None

-

39

Parley

None

-

TMDL

Pierson

25% of residential, agricultural, and vacant/agricultural  load

27

19

Steiger

25% of residential, other, and vacant/agricultural load

19

28

Stone

25% of agricultural load

3

2

Wasserman

25% of agricultural, 15% of residential and vacant/ agricultural load

28

TMDL

Zumbra

10% of residential load

5

8

Direct & Minor

See detail

In direct drainage area, generally a 50-60% reduction in total load to achieve target flow-weighted mean average concentration in runoff.   In minor watersheds, 278 pounds across all subwatersheds.

-

413

Minnehaha Creek

Creek (above Hiawatha)

Lakes TMDL: approx 15% watershed load reduction

-

363

(assumes 10% decrease)

Brownie

None

-

12

Cedar

None

-

56

Isles

None

-

23

Calhoun

None

-

92

Harriet

None

-

28

Nokomis

Lakes TMDL: various activities including P-free fertilizer, street sweeping, operation of weir, neighborhood BMPs

198

TMDL

Diamond

Lakes TMDL: various activities including P-free fertilizer, street sweeping, increased infiltration

124

TMDL

Hiawatha

Lakes TMDL: increasing infiltration, reduce loads, P-free fertilizer in upstream creekshed: general BMPs

1706

TMDL

Creek (below Hiawatha)

None

-

18

 

Powderhorn

Not modeled

-

TMDL

 

 


 

Introduction

One of the objectives of the Third Generation MCWD WRMP is to develop rules for land development which, when implemented, will help lakes in the watershed meet water quality goals.  These rules are to be based on modeling conducted for the HHPLS, i.e., the P-Load watershed loading model and the WiLMS lake response model.  The supporting analysis should show that water bodies are projected to meet water quality goals when the watershed reaches “Ultimate Land Use” (or full development) and MCWD capital projects and other load reductions are implemented.

 

The reader is directed to the HHPLS report for detailed description of the P-Load and WiLMS methodologies.  The descriptions below are provided to elucidate the main implications of the of the model and parameter selections included in the models.

 

This section is also to document the methodology proposed for the assessment and demonstrate its application to a specific subwatershed, namely Stubbs Bay, before application to all subwatersheds.

 

P-Load Background

Water quality modeling of annual runoff, pollutant loading and lake response is carried out using GIS analysis, PLOAD, and WiLMS. 

 

The PLOAD model estimates watershed nutrient export on the basis of runoff volume and assigned event mean concentrations.  These concentrations are assigned on the basis of land use / land cover.  The model is strictly a loading model; as such it does not reflect nutrient loss or substantial recycling in lakes and wetlands.  The model approach is based on a “normal” year and is not suited to model year-to-year variation in watershed runoff or nutrient export.  Its strength is the representation of land use in terms of nutrient export and expectations for future changes in loading.  Its weakness include the broad assumptions of runoff concentrations for use across the watershed and the target annual runoff depths to which the model was calibrated.  These depths, because they range only from 4.2 to 5.0 inches per year, are not likely to represent the actual variation in runoff volume across the watershed.  The runoff in the PLOAD model is not tied to the calibration of the District’s XP-SWMM model, and only partially to the results of the District’s hydrodata program.

 

WiLMS

The Canfield-Bachman in-lake water quality response models applied from WiLMS are a common and accepted approach for assessment of nutrient load effects on lakes in Minnesota.    Typically, the watershed loading was not high enough to explain the in-lake concentrations observed in the lakes.  The additional load needed to calibrate to the average in-lake conditions was called “unknown” in the HHPLS, or attributed to internal loading from the lake sediments.  The lake models would be improved with an independent assessment of internal loads.

 

Implicit Assumptions of the HHPLS P-Load Model

There are three main inputs that to the determination of loads using P-Load:

  • Percent impervious – for determination of runoff coefficient (determined from land use / land cover and associated percent impervious)
  • Annual precipitation depth (average conditions) and the calibration factor, called the “ratio of storms producing runoff”
  • Total phosphorus event mean concentrations (determined from land use / land cover data)

 

These imperviousness assumptions are described below:

Percent Imperviousness Associated with Land Cover

Land Cover Type

Percent Impervious Used in Models

Upland soils with planted, maintained, or cultivated coniferous or deciduous trees, shrubs, or grasses.

1%

Upland soils – row cropland or close grown cropland

1%

Forest, woodland, upland shrubland, savanna, grassland

1%

Hydric soils with planted, maintained, or cultivated coniferous or deciduous trees, shrubs, or grasses.

10%

Hydric soils – row cropland or close grown cropland

10%

Semi-permanently, or permanently flooded shrubland, wet meadow, wet prairie, cattail marsh

10%

Swamp, wet meadow shrub

10%

Floodplain forest, lowland hardwood forest

10%

Open water, including littoral aquatic beds, open water wetlands, floating algae

100%

 

 

Percent Imperviousness Associated with Artificial Surfaces

Percent Impervious Range from MLCCS Codes

Percent Impervious Used in Models

0 - 10%

5%

4 - 10%

7%

11 – 25%

18%

26 – 50%

38%

51 – 75%

63%

76 – 90%

83%

91 – 100%

95.5%

 

The correlation of land use / land cover with phosphorus event mean concentrations was completed using the Minnesota Land Cover Classification system and MCES 2020 Land Use projections (based on comprehensive planning by the municipalities and other data sources).  By their nature, these two data sets are different and do not lend to comparison based on individual areas.  The HHPLS model is based on the 2000 land cover, and altered for 2020 to describe the new development.  The model cannot be explicitly updated to other data sets without repeating the GIS analysis of each subwatershed. 

 

The phosphorus event mean concentrations implicit to the model are listed below.

 

Total Phosphorus Event Mean Concentrations (EMCs)

Associated with Land Cover and Land Use

Land Cover (Undeveloped Land)

TP [mg/L]

Cropland

0.32

Forest/Shrub/Grassland

0.04

Open Water

0.01

Wetlands

0.01 – 0.04

Land Use (Developed Land)

TP [mg/L]

Farmsteads

0.46

Single Family Residential

0.46

Multi-Family Residential

0.32

Vacant/Agricultural

0.32

Airports

0.28

Commercial

0.28

Industrial

0.28

Public Industrial

0.28

Public/Semi Public

0.28

Public/Semi Public Not Developed

0.28

Park and Recreation

0.04

 

There are some special cases of phosphorus EMCs (e.g. wetlands and ponds) described in HHPLS report.

 

Through the combination, or overlay, of the various GIS layers, each subwatershed is parsed into multiple polygons.  These polygons have a wide range of sizes depending upon the source data and intersections of the data.  As an example, in the Classen Creek subwatershed, there are 354 polygons in the 2000 model and 408 in the 2004 model, with an average area of 2.8 and 2.4 acres, respectively.  The range for polygon sizes is from 0.001 to 44 acres.  Median polygon sizes are 0.84 and 0.56 acres, for 2000 and 2020 respectively.

 

Development Impacts as Assessed by P-Load

The main development impacts in much of the MCWD will be seen in the alteration of existing agricultural and forested land to residential land use (e.g., Fig. IV.K.2-2 from the HHPLS, see below).  Much of the new development in the Stubbs Bay watershed will be relatively low density residential, with impervious cover fractions being in the 4 to 25 percent ranges, with practically no additional development in the higher impervious categories.

 

Some examples of the development impacts on phosphorus loading, as predicted by the P-Load model, are summarized in the table below:

 

The per-acre phosphorus loads shown are the maximum predicted by the model.  The specific sub watershed models predict 30 to 80 percent of the volumes and loads.

 

Modeled Impact of Development

Based on the tabular approach described above, the P-Load model developed for the MCWD will predict dramatic increases in load, when considered on a project basis.  For example, if an area with upland soils cultivated for cropland is developed into a low-density single-family development, the model would predict that runoff volume would increase by 250 percent, loads would increase by 417 percent.  Commercial development would increase runoff by 1200 percent and phosphorus loads by 1100 percent. 

 

And all forms of development will far exceed the existing loading from forest areas.

Effect of P-Load Volume Calibration

The P-Load model, as developed for the HHPLS, was calibrated to match annual runoff values as mapped in the Minnesota Hydrology Guide.  Therefore the runoff from all subwatersheds averaged 4.4 to 6 inches annually.  In all cases (i.e. subwatersheds) the model predicted a larger annual runoff volume than determined from the Hydrology Guide.  But there was also substantial variation amongst the subwatersheds; the Pj values ranged from 0.25 to 0.75.  The model, without such calibration would have predicted 1.3 to 3.7 times the runoff and hence load from an otherwise similar area in different subwatersheds. For example, the load from a home site in the Painter Creek watershed would be predicted to be three times that in an identical home site in the Lake Minnetonka Direct subwatershed.  The subwatershed Pj values are summarized at right:

 

 

Methodology for Predicting Ultimate Development Loads

  1. The model results for existing conditions (2000) and 2020 conditions will be used as reported in the HHPLS report and the associated P-Load spreadsheets. 
  2. Ultimate development will be considered as complete development of agricultural lands, and development of one half of the forest lands estimated to exist in 2020. 
  3. The land developed between 2020 and Ultimate Development will be split according to the proportions of new development types predicted for 2000 to 2020.  That is, the Ultimate Development will follow the same proportions of low and high density residential, commercial and other development types as predicted as the development from 2000 to 2020.   

 

 

Prediction of Lake (Water Resource) Load Reduction Requirements

The WiLMS model inputs (annual runoff volume, annual phosphorus loads, internal loads and lake morphometry) developed in the HHPLS will be used in conjunction with the Canfield-Bachmann lake response formulation.  Due to the error in the WiLMS modeling, the internal loads will be reassessed using the corrected model and the average in-lake concentrations developed for the HHPLS/WiLMS.  As in the HHPLS, internal load will be taken as the additional load necessary to explain the observed in-lake concentrations using the corrected model and the P-Load results for 2000.  For each annual runoff volume (existing, 2020, and ultimate) the model will be used to estimate the load (and load reduction) to achieve in-lake phosphorus goals.  The model is relatively insensitive to changes in inflow volume; in some cases a single inflow volume may be sufficient.

 

Methodology for Load Reduction Requirements

For each lake (or stream) with a particular water quality goal, the Canfield-Bachman lake response model will be used to estimate the allowable load that will achieve the target in-lake phosphorus concentration. The lake phosphorus – load relationship will be estimated and plotted for the 2000 and 2020 runoff volume estimates.  The 2000, 2020 and Ultimate Land Use load reductions will be reported without application of BMPs.

 

Evaluation of Performance-Based Rules

The model results will be used to determine, on a watershed by watershed basis:

  • Lake response to development under future conditions without BMPs.
  • Lake response to development under future conditions and current MCWD rules.
  • Proposed framework for revised rules.

 

The results can be used to assess additional reductions needed through capital projects

 

Demonstration Application to Classen Creek and Stubbs Bay  / Maxwell Bay

Determine inputs:

  • Collect loads from Volume IV-Appendix 2 in HHPLS
  • Check total loads and volumes against WiLMS model inputs
    • If necessary track down differences in load.  Differences in volume should be ignored and use the total unless otherwise known to be necessary.  If there are differences explain in a separate sheet called “Notes”.
    • Deduct or otherwise adjust for the lake surface load and volume.

 

For extrapolation to Ultimate Land Use, utilize the following procedure to determine the Ultimate annual volume and the ultimate annual phosphorus load:

  • Collect the 2020 load and volume;
  • Determine the remaining land area to be developed between 2020 and ultimate land use:
    • Assume all MLCCS 1xxx codes are developed in 2020 and not available for further development;
    • Include all 2xxx as remaining cropland, which will become developed at 100 percent;
      • Determine the total of remaining forest and woodland developable area as all 3xxx (forests), 4xxx (woodlands), 5xxx (shrubland) areas with the exception of areas whose description indicates wetland by including the words such as “seasonally flooded”, “wet”, “swamp”, and “semi-permanently flooded”.   In the case of Classen Creek watershed example, this leaves the following codes and descriptions:

 

Code

Description

3215

Maple-basswood forest

3217

Boxelder - green ash disturbed native forest

4213

Disturbed deciduous woodland

5213

Non-native dominated upland shrubland

 

 

(This approach is consistent with the HHPLS, which indicated that in almost every subwatershed, grasslands were shown as not being developed.  In fact, much of the area included in the grassland category is wetland.)

  • Sum up the volumes and loads from these areas from the detailed P-Load calculations in the HHPLS. 
  • Subtract these loads and volumes from the P-Load results for 2020 land use.

 

  • Determine the new loads for the development expected to occur between 2020 and ultimate conditions:      
    • Assume that the new development will occur as residential with either 4 to 10 percent imperviousness or 11 to 25 percent imperviousness.  This is based on the fact that in each of the subwatershed P-Load summaries (as in the example for the Minor Watersheds, which include Classen Creek), practically all of the development expected to occur between now and 2020 is in these two categories. 

 

 

 

  • Determine the ratio between the 4 to 10 percent and 11 to 25 percent categories from the corresponding HHPLS figure as shown above.  The ratios for each subwatershed is determined as follows:

 

Subwatershed

Fraction of New Development in 4 to 10 Percent Impervious Category

Fraction of New Development in 11 to 25 Percent Impervious Category

Painter Creek

46

54

Dutch Lake 

80

20

Langdon Lake 

54

46

Six Mile Creek

25

75

Long Lake Creek

93

7

Gleason Lake Creek

100

0

Schutz Lake 

17

83

Lake Virginia

70

30

Christmas Lake 

100

0

Lake Minnetonka Direct Drainage

79

21

Minor Watersheds

56

44

Minnehaha Creek (Lower Watershed)

100

0

 

  • Determine the load from the future developed areas.  The load from the developed areas will be calculated according to P-Load using 7 and 18 percent imperviousness, and the corresponding 0.46 mg/L phosphorus concentration.  The actual volume of runoff and load will depend on the subwatershed Pj value (See above). 

It is important to note that following exhaustive attempts, it was not possible to correlate the P-Load results with the categories and results presented in the HHPLS report.  This results from the different codes used in the 2000 and 2020 models.  Also, the land areas (i.e. the GIS polygons) are different.  Therefore the categories of land use and cover reported in the figures in the HHPLS report cannot be correlated with the categories in the P-Load model.  Therefore, the procedure outlined above makes use of the 2000 and 2020 P-Load model results as they are, and uses an incremental methodology to assess Ultimate Development loads.