Chatfield Reservoir Phosphorus Budget Jim Saunders and Jamie Anthony WQCD, Standards Unit 13 Dec 2007.

Chatfield Reservoir Phosphorus Budget

Jim Saunders and Jamie AnthonyWQCD, Standards Unit13 Dec 2007

Purpose of Phosphorus Budget

Quantify annual loads associated with inflow sources

Assess relative importance of sources (typical as well as variation)

Determine annual loads as precursor to development of “load translator”

What is Load?

Concentration x Flow => pounds P Daily: mg/L x ft3/s x units factor; get

lbs/day Annual: Σ(daily loads) Can simplify if concentration does

not vary with flow or over time: concentration x Σ(daily flows)

Sampling Frequency Problem

Flow is reported daily for major tribs (South Platte and Plum)

Concentration is measured 15-20 times per year on average (~5% of days)

What is best way to assign concentrations to every day?

Characterize Variation in Concentration

Common for concentration to vary with flow; often lower conc at higher flow If the linkage is strong, can predict

concentration for any flow (regression analysis) Trend over time might be expected in

case of development (or wildfires?) Overlay of patterns associated with flow

and time harder to deal with

South Platte – no connection to flow

0.0

0.2

0.4

0.6

0.8

1.0

1.2

10 100 1000 10000

Waterton Flow, cfs

To

tal

Ph

osp

ho

rus,

mg

/L

Measured Nondetect

More about South Platte

Absence of flow concentration linkage reflects role of upstream reservoirs and intensive flow management

Why not use average? Large range of concentrations; don’t want to eliminate observed variability

Could have subtle flow-concentration link obscured by temporal trend

South Platte TP over Time (MDL problems in some years) Is there a trend over time?

Disentangling Time and Flow

Assume there are underlying links between concentration and both flow and time for the South Platte

Define categories of flow based on quartiles for period of record

Define consecutive blocks of time Assign each day (1987-2006) to a time

block and a flow category based on daily flow

Each phosphorus measurement can be classified in the time-flow matrix based on flow observed on sampling date

Flow Categories for South Platte (20-y)

1

10

100

1000

10000

0% 25% 50% 75% 100%

Flow Percentile

Dai

ly F

low

, cf

s

25th percentile: 32 cfs

75th percentile: 72 cfs

Low HighIntermediate

Classify Phosphorus Measurements by Time and Flow

Divide sampling record into consecutive 5-y time blocks beginning with 1987

Based on sampling date, classify each measured phosphorus concentration according to one of 3 flow categories

Example: TP=0.024 mg/L on 3/29/05; flow was 30 cfs Assign to time block 4 (2002-2006) Assign to low flow category (<32 cfs)

South Platte Phosphorus Measurements

Flow ’87-’91 ’92-’96 ’97-’01 ’02-’06 Total

Low 19 17 12 27 75

Mid 35 53 39 37 164

High 39 40 33 19 131

Total 93 110 84 83 370

Compare Years using Phosphorus Measured during Intermediate Flows

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0.050

2002 2003 2004 2005 2006

Year

Ph

os

ph

oru

s C

on

ce

ntr

ati

on

(m

g/L

)

Overview of Phosphorus Comparisons

No difference in concentration across years within intermediate or high flow categories in any of the 5-y blocks

Sample size too small for same comparison in low flow category, but will assume no difference

Safe to lump phosphorus data across years within flow categories within time blocks

Are there patterns over time within flow categories?

Random Sampling Methodology

Flow and time disentangled as much as practicable (3 flow x 4 time units)

Load strategy: assign concentration to every day based on flow in South Platte

Preserve variation observed in concentration data (i.e., don’t use avg)

In each time block, each measured concentration in a flow category is equally valid for every date that falls in that category (i.e., can select at random)

Random Sampling Example 1994

1994 flows; 31%:49%:21%

1994 TP; 7:8:7

Annual Loads with Replication 365 daily loads, summed for annual load Can repeat as often as you want; we did 100 reps

Plum Creek TP over Time Is there a trend? Not obvious

0

0.1

0.2

0.3

0.4

0.5

0.6

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

To

tal P

ho

sph

oru

s (m

g/L

)

Flow Categories in Plum Creek (20-y)

25th percentile: 4.4 cfs

75th percentile: 24 cfs

Low HighIntermediate

Plum Creek Phosphorus Data

Flow ’87-’91 ’92-’96 ’97-’01 ’02-’06 Total

Low 14 4 10 10 38

Mid 35 35 34 43 147

High 36 46 31 10 123

Total 85 85 75 63 308

Flow Patterns for Plum Creek?Yes, but can ignore time

Plum Creek: Use Flow Deciles

No apparent temporal trend Collapse time blocks into one “Noisy” linkage to flow Subdivide flows into 10 flow categories

(deciles) All deciles (except lowest flow) have at

least 24 concentration measurements Lowest decile – only one concentration; all

zero flow days

Plum Creek Annual Loads with Reps

Note magnitude comparable to Platte

More Phosphorus Sources

Direct Precipitation Monthly avg lake area * Kassler gage …monthly AF * [TP] Set [TP] to 0.087 mg/L (Clean Lakes

study) in all months Alluvial

Estimated annual inflow * [TP] Set [TP] to 0.010 mg/L (monitoring

data) in all years

Phosphorus from Ungaged Areas

Limited data for ungaged tribs (1.4% of basin area); no data for direct runoff (0.4%)

Assume TP yield similar to Plum Creek watershed; i.e., scale up by water yield

Total Phosphorus Load

45%

45%

1%

7%2%

South Platte River

Plum Creek

Alluvium

Ungaged Runoff

Precipitation

Compare Loads

Old vs. new load method; compare to equivalence line

SP shows strong bias (new>old)

Plum similar Assumptions behind

methods likely different

New method benefits from review of 20-y record

South Platte

Plum Creek

Conclusions and Comments

Random sampling methodology provides robust and flexible approach to load estimates for gaged inflows

Gaged inflows contribute about 90% of annual load; even split between SP and Plum (but much variation among years)

Small contribution likely for precip and alluvium; apply simple method

Some concern about under-estimating Plum due to issues with water budget

Expect to make spreadsheets available next month