| I, WILLIAM W. WALKER, JR., PH.D., declare:
1. I am an independent consultant. I received S.B. and S.M.
degrees
in Chemical Engineering from Massachusetts
Institute of Technology
in 1971 and a Ph.D. in Environmental
Engineering from Harvard
University in 1977.
2. I have 20 years of experience in analysis of data and
modeling of
environmental effects of water and waste
management practices in
surface water bodies.
3. Before becoming an independent consultant, I was an
Environmental
Engineer vith Meta Systems, Inc., from 1975 to
1980, and with
Process Research, Inc., from 1972 to 1975. My
work in those
positions involved modeling of eutrophication
and other water-
quality processes in rivers, lakes, and
estuaries; evaluation of
urban and agricultural nonpoint source
pollution; and lake
restoration.
2
4. My current clients include the City of St. Paul Water
Utility,
USEPA's Narragansett Bay Project, and State of
Oklahoma, among
others. Previous clients have included the U.S.
Army Corps of
Engineers, U.S. Environmental Protection
Agency, City of Baltimore
Water Supply, Minnesota Pollution Control
Agency, Vermont
Department of Water Resources, and New York
State Department of
Environmental Conservation, among others. My
work for these
clients has focused on data analysis and
modeling to help identify,
evaluate, and solve water quality problems,
with emphases on
problems relating to eutrophication and
nonpoint source pollution.
5. Between 1978 and 1989, I conducted an extensive research
project on
reservoir eutrophication for the U.S. Army
Corps of Engineers
Waterways Experiment Station. This project grew
out of my doctoral
thesis research. The work involved compiling
and analyzing a
nationwide data base on morphometry, hydrology,
and water quality
in Corps of Engineers reservoirs. I used the
data to develop and
test empirical models for predicting
eutrophication and related
water quality conditions in reservoirs. These
models predict
reservoir nutrient and algae concentrations as
a function of inflow
nutrient concentrations, flow, and reservoir
dimensions (surface
area, volume, depth). The final phase of the
project involved
development of computer software to assist
Corps staff and other
model users in analyzing data from river and
reservoir monitoring
stations and in applying the model.
6. In recognition of research conducted for the Corps of
Engineers and
other agencies, I received a "Technical
Excellence Award for
Outstanding Research in Lake Restoration,
Protection, and
Management" from the North American Lake
Management Society in
1988.
7. I have conducted an independent analysis of water quality
data
collected at inflow points to Everglades
National Park (ENP)
between 1977 and 1989. The data were derived
from a routine
3
biweekly monitoring program operated by
South Florida Water
Management District since 1977. The objective
of the analysis was
to determine whether trends in nutrients or
other water quality
components could be identified during this
monitoring period.
Results are described in the attached report
entitled "Water
Quality Trends at Inflows to Everglades
National Park". Figures
referenced below are contained in the report.
8. Results indicate that the quality of water entering
Everglades
National Park declined during the 1977-1989
period. Increasing
trends in total phosphorus concentration and
decreasing trends in
the total nitrogen to phosphorus ratio were
observed at most ENP
inflow points. In other words, over time there
was a statistically
significant increase in phosphorus
concentration and a
statistically significant decrease in the total
nitrogen to
phosphorus ratio in the water entering ENP.
Total phosphorus
concentrations increased at rates ranging from
4% to 21% per year.
These nutrient trends are classic symptoms of
eutrophication, a
process which could alter the unique
oligotrophic character of ENP
marshes. Trends in other water quality
components were detected at
much lower frequencies and were of less
importance with respect to
potential ecological impacts.
9. The analysis included data for 20 water quality components,
including nutrients (phosphorus and nitrogen
compounds), field
measurements (dissolved oxygen, temperature,
pH, conductivity),
inorganic species (chloride, etc.) and optical
properties (color,
turbidity). To permit consideration of
hydrologic variations,
concentration measurements were paired with
daily flow, water
elevation, and rainfall data collected by South
Florida Water
Management District and U.S. Geological Survey.
Mass transport
rates (weight per unit time) of total
phosphorus and total nitrogen
were also calculated from flow and
concentration measurements and
tested for trends.
4
10. The analysis considered data from seven ENP inflow stations
(Figure
1). Five of these stations were located in
Shark River Slough
(S12A, S12B, S12C, S12D, and S333), one in
Taylor Slough (S332),
and one in ENP's Coastal Basin (S18C). In
addition, two composite
stations were analyzed to reflect combined
discharges to Shark
River Slough (S12's and total). The sampling
period was 1977-1989
for Shark River Slough stations and 1983-1989
for the remaining
stations.
11. As illustrated in Figure 3, three time series (data sets) were
tested for each station and water quality
component: (A) all
concentration data; (B) a subset of
concentration data collected on
days during which there was appreciable flow
into ENP at the
sampling station; and (C) the data analyzed in
a manner which
accounted for variations in antecedent rainfall
(i.e., rainfall
prior to each sampling date) and water surface
elevation. The
objective of the third series of tests was to
distinguish water
quality changes that were related to hydrologic
variations (e.g.,
wet vs. dry periods) from longterm trends.
12. I applied a statistical technique called the "Seasonal Kendall
Test" to determine the likelihood that an
underlying trend existed
in each tested data set. This technique was
originally developed
by the U.S. Geological Survey and used in a
nationwide study of
phosphorus trends in U.S. river basins.
Desirable features of the
Seasonal Kendall Test include that it is
nonparametric (does not
assume a particular underlying probability
distribution for the
data), it can be applied to data sets
containing observations that
are missing or below detection limits, and it
accounts for seasonal
variations. Because of these desirable
features, the test has been
widely recommended and applied to water quality
data. For example,
the Seasonal Kendall Test was used in a study
of water quality data
collected between 1978 and 1982 in Everglades
Water Conservation
Area 3A; the study was conducted jointly by the
U.S. Geological
5
Survey, National Park Service, and South
Florida Water Management
District.
13. The Seasonal Kendall Test estimates the probability that a trend
does not exist in the data set being analyzed.
This probability is
termed the "significance level" and
is denoted by the mathematical
symbol "p". Lover values of the
significance level indicate that a
trend is more likely. In summarizing and
discussing results for
each tested data set, the presence of a trend
was concluded when
the significance level was less than .10.
According to this
convention (also used in previous water quality
studies), the
existence of a trend is concluded only when the
risk of error is
less than 10%, or when the "confidence
level" in the conclusion
exceeds 90%. This is the definition of a
"detected trend".
14. I estimated trend magnitude for each data set using a procedure
developed by the U.S. Geological Survey. Trend
magnitude expresses
the median rate of increase or decrease during
the sampling period
in units of percent per year. No assumptions
were made regarding
underlying trend shapes (e.g., linear,
exponential, step change at
a specific date).
15. Results for nutrient species were dramatically different from
results for other water quality components, in
that nutrient trends
were detected at much higher frequencies
(Figure 9). Increasing
trends in total phosphorus concentration were
detected at 8 out of
9 ENP inflow stations examined. When the data
were adjusted to
account for hydrologic variations, increasing
trends were detected
at 7 out of 9 stations (all but S333 and S18C).
Trend magnitudes
ranged from 4% per year at S12D to 21% per year
at S332.
Decreasing trends in total nitrogen
concentrations were detected at
5 Shark River Slough stations. Trend magnitudes
ranged from -4%
per year (S12D) to -3% per year (combined
discharge to Shark River
Slough). Decreasing trends in the total
nitrogen to phosphorus
6
ratio were detected at 7 out of 9 stations.
Trend magnitudes
ranged from -7% per year (S12B) to -15% per
year (S332).
16. Increasing trends in phosphorus concentration were detected for the
combined discharge through the S12 gates into
Shark River Slough,
the inflow with the longest period of record
and historically of
greatest importance to ENP. Estimates of trend
magnitude (percent
increase in phosphorus concentration per year)
ranged from 5.3 to
7.0% per year and were insensitive to
adjustment of the data to
account for variations in rainfall and water
elevation.
Significance levels for the S12 total
phosphorus trend tests ranged
from .06 to .009; these supported conclusion of
an increasing trend
at confidence levels ranging from 94% to 99.1%.
17. Impacts on biological communities may be related to changes in the
frequencies of extreme phosphorus
concentrations, as well as to
changes in central tendency (i.e., median
concentration). The
phosphorus trend at the S12's was accompanied
by increases in the
frequencies of phosphorus concentrations
exceeding 0.01, 0.02, and
0.03 milligrams per liter, as shown in Figure
7. For example, the
frequency of concentrations exceeding .03
mg/liter increased from
6% in the first 5 years of monitoring
(1977-1982) to 15% in the
last 5 years of monitoring (1984-1989). These
increases in
exceedence frequencies reflect the underlying
trend in phosphorus
concentration.
18. The analysis of S12 total phosphorus data was modified in several
ways to assess the sensitivity of the results
to the methods used,
to the hydrologic variables considered
(rainfall, flow, water
elevation), and data subsets. This was done to
determine the
extent to which results and conclusions were
influenced by the
assumptions, methods, and data used in the
analysis. Regression
models can be used to analyze the strength of
relationships between
water quality parameters and hydrologic
variables. Regression
models were developed considering rainfall,
flow, and elevation
7
separately and combined, using different lag
periods to represent
the effects of these variables, and using
alternative rainfall
averaging methods. The analysis was repeated
excluding data
between January 1 and September 15, 1985
(period of a pronounced
phosphorus spike at the S12's shown in
Figure 4). The data were
split in three different ways (wet season
versus dry season, low
flow versus high flow, and low elevation versus
high elevation) and
each subset was analyzed for trend separately.
Test results were
generally insensitive to these factors; this
strongly supported the
conclusion that increases in total phosphorus
concentration
occurred at the S12's between 1977 and 1989. In
fact, the
sensitivity analysis indicated that the
increasing trend in total
phosphorus concentration at the S12's was more
distinct under high
flow conditions, which accounted for 87% of the
total water volume
discharged to ENP through the S12's during this
period.
19. An increasing trend in the rate of phosphorus transport (kilograms
per day) into Everglades National Park was
detected for the Taylor
Slough basin. When adjusted for hydrologic
variations, increasing
trends in phosphorus transport were detected in
all three ENP
basins (Shark River Slough, Taylor Slough, and
Coastal) and a
decreasing trend in total nitrogen transport
was detected in Shark
River Slough. Trend magnitudes were similar to
those estimated for
concentration.
20. Trends were also detected in the remaining water quality components
(other than nutrients), but at a much lower
frequency. There was a
disproportionately large number of significant
results among all of
the water quality variables--more than could be
expected by chance
alone. From a water quality management
perspective, the increasing
trends in total phosphorus (Figure
10) and
decreasing trends in the
total nitrogen to phosphorus ratio
(Figure 11)
were most important
because they are classic symptoms of
eutrophication.
8
21. Increasing phosphorus trends at ENT inflow stations between 1977
and 1989 were detected in three forms:
concentration, transport
rate, and frequencies of values exceeding 0.01,
0.02, and 0.03
mg/liter. Thus, biological communities in
Everglades National Park
were subjected to effects of increased average
phosphorus
concentration, increased mass of phosphorus,
and increased
frequency of concentrations above benchmark
levels during this
period. Although decreasing trends in total
nitrogen were detected
at some Shark River Slough stations, the
phosphorus trends are much
more important because phosphorus, not
nitrogen, controls aquatic
productivity in the Everglades ecosystem.
22. The Everglades Water Conservation Areas (WCA's) serve as shallow
wetland reservoirs in supplying water to ENP
and in meeting other
South Florida water needs. The input/output
concept is fundamental
to reservoir (or lake) management. According to
this concept,
reservoir outflow water quality is influenced
by reservoir inflow
water quality. Increases or decreases in
reservoir nutrient
loadings or inflow concentrations often lead to
increases or
decreases in nutrient concentrations measured
in reservoir
outflows. Reservoirs vary considerably with
respect to the degrees
and time scales of input/output response.
Additional phosphorus
loadings were intentionally diverted to the
WCA's during the 1977-
1989 period as a result of the Interim Action
Plan. Increases in
the intensity of watershed land use could also
have increased
phosphorus loadings to the WCA's during this
period. The
increasing phosphorus trends detected at ENP
inflows between 1977
and 1989 were not explained by seasonal factors
or by variations in
rainfall, water elevation, or flow. It is
possible that these
trends reflected increases in phosphorus
loadings to the Water
Conservation Areas during this period.
23. The increasing total phosphorus concentrations and decreasing N/P
ratios detected at ENP inflows are symptoms of
eutrophication, a
process which generally leads to undesirable
conditions, including
9
shifts in aquatic species, low dissolved
oxygen concentrations, and
diminished wildlife habitat. This process must
be avoided if the
unique water quality and ecology of the
Everglades National Park
marshes are to be preserved.
I declare under penalty of perjury that the foregoing is true and
correct to the best of my knowledge and belief.
Dated this 17th day of September, 1990.
_______________________
Dr. William W. Walker, Jr.
|
