I-48_18-002 WQI DO modifications_2021.11.23-Final

CITY OF DURHAM
Department of Public Works
Stormwater & GIS Services
CITY OF
DURHAM
Water Quality Report #18-002
November 23, 2021
Modifications to Water Quality Index scoring of dissolved oxygen data to
remove natural seasonal trends
1.0 Introduction
This report looks at methods to remove natural seasonal trends in dissolved oxygen (DO) data from Water
Quality Index (WQI) calculations. The Water Quality Unit (WQU) in the Public Works Department's Stormwater
and GIS Services Division computes a WQI score at each monitored site to represent the overall general health
of Durham's streams. These WQI scores are calculated monthly. These monthly scores are also combined to
determine annual WQI scores for each watershed and an overall City WQI score. These WQI scores are based on
data collected by the WQU, including DO, turbidity, metals, and other parameters measured in the stream. Use
of DO in the WQI is typical for these types of water quality indices (e.g., Stream Use Support Index [Charlotte -
Mecklenburg], Anacostia River Report Card) and has associated benefits and drawbacks. Low DO concentrations
measured in a stream can be helpful in indicating the presence of pollution like wastewater, ammonia, and
nutrients. Conversely, high DO values usually indicate good water quality.
One issue with using DO in WQI-type calculations is that it varies due to both human -causes and natural factors
that change with the seasons. Regarding these natural factors, DO values are naturally lower in warmer water.
Thus, in Durham's streams, DO is typically lower in the hotter summer months and higher in the cooler winter
months. Low -flow conditions in streams, which commonly occur during dry summer months, can also lead to
lower DO values due to stagnation and low aeration. Since DO is used to determine the WQI, the WQI is also
impacted by these natural factors that change with the season. One of the uses of the WQI is to identify areas
affected by pollution; however, natural seasonal trends muddle the effectiveness of the WQI to show pollution
hot spots. The goal of this report is to look at ways to reduce the impact of natural seasonal trends in DO data
on the WQI calculations.
2.0 Methods - Evaluation of Three Options
In this report, three options to modify the DO part of the WQI were evaluated and compared to the current
approach. In the calculations associated with the current WQI (termed Current Option in this report), DO alters
the WQI score by adding a modifier value to the base WQI score calculated for each site for each month. These
WQI modifier values currently used are shown in Table 1.

Table 1. Current Option that is used in the WQI to determine WQI modifier values based on DO concentration data.
DO Concentration (mg/L) WQI Modifier Value
>7 +2
5-7 0
4-5 -15
<4 -24
Three different alternatives to the Current Option (Table 1) were developed and tested to remove seasonal
trends from WQI scores. These three alternative options included:
• Option 1: Replacement of WQI modifier values based on DO concentration ranges with WQI
modifier values based on DO saturation ranges. For Option 1, the scheme in Table 2 was used.
• Option 2: Averaging the WQI modifier values to be based on both DO concentration ranges and DO
saturation ranges. This option is simply the average of WQI modifier values determined in the
Current Option and Option 1.
• Option 3: Using DO concentration data like in the Current Option but, altering the WQI modifier
values based on the month. For example, a DO measurement of 6 mg/L would have a higher WQI
modifier value in a hotter month as compared to a colder month. For Option 3, the scheme in Table
3 was used.
Table 2. Option 1 scheme used to incorporate DO saturation values, rather than DO concentration values, into monthly WQI
calculations.
DO Saturation (%) WQI Modifier Value
>70 +2
70-55 0
40-55 -15
<40 -24
Table 3. Option 3 scheme with WQI modifier values based on DO concentration that vary by month.
DO
WQI Modifier Value by
month
Concentration
Range (mg/L)
Jan.
Feb.
Mar.
Apr.
May
Jun.
Jul.
Aug.
Sept.
Oct.
Nov.
Dec.
>7
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
+2.0
5-7
0.0
0.0
0.0
+1.0
+1.0
+1.0
+1.0
+1.0
+1.0
0.0
0.0
0.0
4-5
-1 5.0
-1 3.0
-11.0
-9.0
-7.0
-5.0
-5.0
-7.0
-9.0
-11.0
-1 3.0
-1 5.0
<4
-24.0
-21.2
-18.4
-15.6
-12.8
-10.0
-10.0
-12.8
-15.6
-18.4
-21.2
-24.0
To evaluate these three proposed options, DO data collected in 2019 were analyzed. These data included typical
cooler winter months, hotter summer months, and a period of severe drought during September and October.
Monthly WQI modifier values were determined for each monitoring site using the Current Option and Options 1-
3.0 Results
3.1 Comparison of the Current Option with Options 1, 2, and 3
DO concentrations measured in Durham's streams during 2019 showed the typical seasonal trend (Figure 1).
Figure 1 shows the total number of sites that fell into buckets used to calculate the WQI modifier values using
2

the Current Option (i.e., >7 mg/L, 5-7 mg/L, 4-5 mg/L, and <4 mg/Q. Seasonal trends can be seen with more sites
having lower DO concentrations during hotter months and more sites having relatively higher DO concentrations
during cooler months (i.e., dark blue line is lower in May -October and higher in December -March). DO
concentrations at all sites declined significantly during September and October when severe drought conditions
were present; the average DO concentration measured in the field during these two months was 4.9 mg/L.
W
,,, 20
m
in
w 15
O
L
J0 10
E
z 5
0
1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure 1. The total number of sites monitored each month during 2019 that fell into four different DO concentration buckets (i.e., >7
mg/L, 5-7 mg/L, 4-5 mg/L, and <4 mg/L) used to calculate the WQI modifier values in the Current Option.
DO modifier values determined from 2019 data using the Current Option and Options 1-3 reflected the DO
concentrations showed above (Figure 2). In Figure 2 the average DO modifier value for the Current Option and
Options 1-3 are shown to decline during the hotter months (May -October) and were relatively higher during
cooler months (December -March). Average WQI modifiers for all options declined significantly during
September and October when severe drought conditions were present.
4.0
L 2.0
:= 0.0
O v' -2.0
E —
i -4.0
-6.0
GJ OJ
eon -8.0
L M
j } 10.0
a
-12.0
$Current Option
{ Option 1 - Replace with DO sat
$Option 2 - 50%DO 50% DO Sat
$Option 3
1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure 2. Average WQI modifier value for all sites by month for the Current Option and Options 1-3.
The largest seasonal trend can be seen in the Current Option with the dark blue line having the lowest average
WQI modifier value during hotter months (May -October), but similar values, as compared to Options 1-3, during
cooler months (December -March). Options 1-3 all showed a seasonal trend to varying degrees. The least marked
seasonal trend was in Option 3 (red line). This is shown with the red line always being the highest on Figure 2.
3

Option 1 had the next smallest seasonal trend. The seasonal trend observed for Option 2 fell between Option 1
and the Current Option. This is logical as Option 2 was the average of Option 1 and the Current Option. Options
1-3 all showed a marked decrease in the WQI modifier value during the drought from September to October,
though to a lesser extent as compared to the Current Option. The average difference between WQI modifier
values for the Current Option and Option 3 during May -August was 2.7. Based on results from Figure 2, Option 3
was examined in more detail to assess this approach to replace the Current Option in calculating WQI scores.
3.2 Detailed Comparison between the Current Option and Option 3
Both the Current Option and Option 3 used DO concentration buckets of <4 mg/L, 4-5 mg/L, 5-7 mg/L, and >7
mg/L to determine a WQI modifier value. The key difference between these options is the magnitude of the WQI
modifier values resulting from a measurement falling into a concentration bucket. Figure 3 and Figure 4 show
how 2019 monitoring data resulted in the WQI modifier values under the Current Option and Option 3,
respectively. While WQI modifier values only consisted of four numbers, +2, 0, -15 and -24, for the Current
Option, fifteen different values were used for Option 3. Thus, ranges of WQI modifier value, as opposed +2, 0, -
15 and -24, were graphed for Option 3 in Figure 4.
25
0 20
0
in I
1 2 3 4 5
6 7 8 9 10 11 12
Month
Figure 3. The total number of sites receiving a WQI modifier value of 2, 0, -15, and -24 under the Current Option for 2019 data.
25
N 20
a
15
O
L
CD
-0 10
E
z
5
0
on I I I I
1 2 3 4 5
■ 2 to 0.9
0.9 to -7.5
1-7.5 to -19.5
1-19.5 to -24
6 7 8 9 10 11 12
Month
Figure 4. The total number of sites receiving a WQI modifier values between 0.9-2.01 -7.5-0.9, -19.5= 7.5, and -19.5= 24 under Option 3
for 2019 data.
il

Differences in how WQI modifier values from the Current Option and Option 3 are influenced by seasonal trends
in DO data are clearly distinguishable. For the Current Option, the number of sites that score a +2 (dark blue
line) dramatically declined during May through August with most of those sites switching over to a score of 0
(light blue line; Figure 3). A few sites score much lower; an average of two sites per month scored -24 during the
hotter months (May -August). In contrast, for Option 3, most sites remained in the range of +0.9 to +2 for the
months of May through August (Figure 4). About two sites per month scored in the -19.5 to -7.5 range during
May through August and no site scored -24 to -19.5 range during that period. During these hotter months, a
majority of sites sampled had DO measurements of between 5-7 mg/L, which is above the state instantaneous
water quality standard. The generally positive WQI modifier values assigned in Option 3 during these hotter
months reflects acceptable water quality under harsh environmental conditions.
Measurements of DO concentration were below 5 mg/L for several sites throughout 2019 (Table 4). All of these
measurements were assigned WQI modifier values of -24 or -15 in the Current Option. Option 3 assigned WQI
modifier values of -24 to -5 depending on the sample date and DO concentration measured. For these low DO
values, the magnitude of most WQI modifier values increased from the Current Option to Option 3, especially
during the hotter months. However, for all DO concentration measurements of less than 5 mg/L, the WQI
modifier values remained negative in value. This means that DO measurements below the state instantaneous
standard of 5 mg/L decrease the WQI score at a site under Option 3. The extreme drought conditions present in
September and October caused low DO concentrations at many stations; these measurements were assigned
very low WQI modifier values (-24 to -15.6) for both the Current Option and Option 3 (gray cells in Table 4).
Under Option 3, the only DO measurement that was assigned the lowest WQI modifier was a measurement of
1.2 mg/L in January at NH1.7SCTA, a site of for low DO concentrations persisting throughout the year.
Table 4. DO measurements collected in 2019 with concentrations below 5 mg/L and WQI modifier values associated with these
measurements for the Current Option and Option 3.
WQI Modifier Value
Date
Site
DO Concentration
DO Saturation
Current
(mg/L)
N
Option 3
Option
NH1.7SCTA
1 /22/2019
1.2
10
-24.0
-24.0
NH3.3SC
1/22/2019
4.7
34
-15.0
-15.0
EL1.9EC
4/2/2019
0.2
2
-24.0
-15.6
EL5.5GC
4/2/2019
0.4
3
-24.0
-15.6
EL7.1 EC
4/2/2019
3.5
31
-24.0
-15.6
EL7.1 SEC
4/2/2019
1.4
12
-24.0
-15.6
NH1.7SCTA
5/21/2019
3.3
37
-24.0
-12.8
NH4.4SCTD
5/21/2019
3.8
43
-24.0
-12.8
EL5.5GC
6/5/2019
3.8
42
-24.0
-10.0
EL7.1 SEC
6/5/2019
2.8
33
-24.0
-10.0
EL8.1 GC
6/5/2019
3.5
40
-24.0
-10.0
LL3.4LLC
6/11/2019
4.4
50
-15.0
-5.0
NH1.7SCTA
6/18/2019
4.4
50
-15.0
-5.0
EL8.1 GC
7/9/2019
3.4
41
-24.0
-10.0
LL3.4LLC
7/16/2019
1.5
17
-24.0
-10.0
EL8.1 GC
8/6/2019
4.1
49
-15.0
-7.0
LL3.4LLC
8/13/2019
1.6
19
-24.0
-12.8
NH1.7SCTA
8/27/2019
4.2
48
-15.0
-7.0
LL3.4LLC
9/10/2019
1.5
18
-24.0
-15.6
NH1.7SCTA
9/17/2019
3.0
35
-24.0
-15.6
NH2.3MC
9/17/2019
3.5
41
-24.0
-15.6
NH3.ONHC
9/17/2019
3.5
41
-24.0
-15.6

NH4.4SCTD
9/17/2019
3.8
44
-24.0
-15.6
EL7.1SEC
10/1/2019
3.0
36
-24.0
-18.4
EL8.1GC
10/1/2019
1.5
17
-24.0
-18.4
EL8.5SEC
10/1/2019
3.3
38
-24.0
-18.4
NH1.7SCTA
10/15/2019
3.7
39
-24.0
-18.4
NH3.ONHC
10/15/2019
1.6
17
-24.0
-18.4
Measurements in gray were collected while drought conditions were present in September and October 2019
4.0 Recommendation
Seasonal trends in the Current Option of the WQI are caused, in part, by natural factors affecting DO
concentrations in the stream. These seasonal trends in the Current Option makes the separation of natural
trends and human -caused issues (e.g., pollution discharge) in WQI scores impossible to separate. An approach to
remove this seasonal variation, Option 3, successfully removed much of the natural seasonal trend typically
observed in WQI scores. While Option 3 did remove the typical seasonal trend, Option 3 did still reflect poor DO
conditions that were caused by extreme drought conditions or poor DO conditions (e.g., NH1.7SCTA site) when
concentrations were below state standards.
The effect of implementing Option 3 will likely improve WQI scores at several sites, especially during hotter
months as compared to using the Current Option. The intent of this change is not artificially inflate WQI scores
at all sites but rather, modify the WQI to provide its users more useful data to identify and address problematic
conditions in Durham's streams. The proposed changes will make it so low DO concentrations, that are caused
by natural factors, do not mask other water quality issues by biasing the WQI scores to lower values during
hotter months.
Based on the enclosed analysis, it is recommended that the WQU update the WQI with the methods used in
Option 3. This includes coding Option 3 methods into the Web Portal, where the WQI is currently calculated.
This may be performed in the near future when the Web Portal is fixed to address ongoing issues. Following the
update of the WQI code with Option 3 methods, the data collected by the WQU dating back to 2004 should be
recalculated using the new WQI method and assessed for potential trends that indicate pollution sources to
Durham's streams. It is also recommended that code to calculate the WQI including Option 3 should also exist in
a separate platform from the Web Portal. This could include software programs like Microsoft Excel or JMP. In
addition to providing redundancy for calculating the WQI in case the Web Portal is inoperable, this approach
would also provide a quality assurance check for the Web Portal.
no