Evaluating the influence of lake morphology, trophic status and diagenesis on geochemical profiles in lake sediments

Abstract

Recent geochemical studies provide evidence that changes in vertical distributions of nutrients in lake sediments are driven by anthropogenic activities, based primarily on trends of increasing concentrations in upper sediment layers. However, the present study shows that vertical concentration profiles of C, N and P in lake sediments can be higher in the upper, most recently deposited sediment strata, driven largely by natural diagenetic processes and not eutrophication alone. Sediment cores from 14 different lakes in New Zealand and China were examined ranging from oligotrophic to highly eutrophic and shallow to deep, and it was found that the shape of vertical profiles of total P, a key nutrient for lake productivity, can be similar in sediments across gradients of widely differing trophic status. Empirical and mechanistic diagenesis steady state profile models were derived and applied to describe the vertical distribution of C, N and P in the sediments. These models, which focus on large scale temporal (decades) and spatial (up to 35 cm in the vertical) processes, revealed that density-differentiated burial and biodiffusive mixing, were strongly correlated with vertical concentration gradients of sediment C, N and P content, whereas lake trophic status was not. A sensitivity analysis of parameters included in the diagenetic model further showed that the processes including flux of organic matter to the sediment–water interface, burial (net sedimentation), breakdown of organic matter and biodiffusion all can significantly influence the vertical distribution of sediment P content. It was concluded that geochemical studies attempting to evaluate drivers of the vertical distribution of sediment C, N and P content in lake sediments should also account for the natural diagenetic drivers of vertical concentration gradients, assisted with application of similar models to those presented in this study. This would include quantification of key sediment diagenesis model parameters to separate out the influence of anthropogenic activities.

Full text

Page 1

1
1
2
3
4
5
6
7
8
9
10
11
12
13
Evaluating the influence of lake morphology, trophic status and diagenesis
on geochemical profiles in lake sediments

Dennis Trolle1,2*, David P. Hamilton1 and Conrad Pilditch1


1 Department of Biological Sciences, University of Waikato, Private Bag 3105, 3240
Hamilton, New Zealand
2 Department of Freshwater Ecology, National Environmental Research Institute, University
of Aarhus, Vejlsøvej 25, PO Box 314, 8600 Silkeborg, Denmark

* Corresponding author: Dennis Trolle (dtr@dmu.dk)

Page 2

2
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Abstract
Recent geochemical studies provide evidence that changes in vertical distributions of
nutrients in lake sediments are driven by anthropogenic activities, based primarily on trends
of increasing concentrations in upper sediment layers. However, we show that vertical
concentration profiles of carbon (C), nitrogen (N) and phosphorus (P) in lake sediments can
be higher in the upper, most recently deposited sediment strata, driven largely by natural
diagenetic processes and not eutrophication alone. We examined sediment cores from 14
different lakes in New Zealand and China ranging from oligotrophic to highly eutrophic and
shallow to deep, and found that the shape of vertical profiles of total P, a key nutrient for lake
productivity, can be similar in sediments across gradients of widely differing trophic status.
We derived and applied empirical and mechanistic diagenesis steady state profile models to
describe the vertical distribution of C, N and P in the sediments. These models, which focus
on large scale temporal (decades) and spatial (up to 35 cm in the vertical) processes, revealed
that density-differentiated burial and biodiffusive mixing, were strongly correlated with
vertical concentration gradients of sediment C, N and P content, whereas lake trophic status
was not. A sensitivity analysis of parameters included in the diagenetic model further showed
that the processes including flux of organic matter to the sediment-water interface, burial (net
sedimentation), breakdown of organic matter and biodiffusion all significantly can influence
the vertical distribution of sediment P content. We conclude that geochemical studies
attempting to evaluate drivers of the vertical distribution of sediment C, N, and P content in
lake sediments should also account for the natural diagenetic drivers of vertical concentration
gradients, assisted with application of similar models to those presented in this study. This
would include quantification of key sediment diagenesis model parameters to separate out the
influence of anthropogenic activities.

Page 3

3
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
Introduction
Internal nutrient loading can directly affect lake trophic status and substantially delay lake
ecosystem responses to reduced external loading (Marsden, 1989; Søndergaard et al., 2003;
Jeppesen et al., 2005). The size and availability of the nutrient pool in the bottom sediments
are therefore of critical importance in understanding how lake ecosystems will respond to
changes in external loading (Nürnberg, 1984; Van der Molen et al., 1998; Spears et al., 2007)
or changes in climate (Jeppesen et al., 2007).

Recognizing that lake sediments can provide information about historical changes in lakes,
the vertical distribution of nutrients in lake sediments is often used to describe how lake
trophic state may have changed through geological time and from recent human activities
(Selig et al., 2007; Xu and Jaffé, 2009). Hence, several studies have used observed profiles of
phosphorus and organic nitrogen and carbon to quantify temporal variations in sediment
nutrient accumulation rates, and infer changes in lake trophic status (Schelske and Hodell,
1995; Hambright et al., 2004; Smoak and Swarzenski, 2004). Only a few studies, however,
have compared sediment geochemical profiles collected from a range of lakes of different
trophic states (Bortleson and Lee, 1974; Søndergaard et al., 1996). These studies focused only
on shallow lakes and did not examine relationships between vertical geochemical profiles and
lake trophic state. Concurrently, both laboratory and field studies have indicated that various
phosphorus species may migrate vertically through the sediments (Carignan and Flett, 1981;
Søndergaard et al., 1996). It is also well known that organic species of phosphorus, nitrogen
and carbon will undergo a natural decay with time (Reitzel et al., 2007), thereby generating
naturally lower concentrations in the deeper and older sediments. Although often assumed
negligible (e.g., Smoak and Swarzenski, 2004), these natural processes may be similarly

Page 4

4
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
important compared to changes in the flux of nutrients to the sediment-water interface
resulting from changes in lake trophic state, in terms of their effect in creating vertical
concentration gradients of sediment nutrient content. We therefore hypothesise that natural
processes should be accounted for in relationships between nutrient concentration profiles in
the sediments and historical changes in lake trophic state. In addition, several studies have
shown that that the surficial sediment concentrations of both phosphorus and nitrogen cannot
readily be related to lake trophic state (McColl, 1977; Håkanson, 1984; Trolle et al., 2008).
Consequently, there is a large degree of uncertainty as to whether, or how strongly, specific
profile properties (e.g., the shape of the vertical concentration profiles) are related to lake
trophic state.

Diagenetic models that describe the vertical distribution of various geochemical elements may
help to quantify the importance of factors such as trophic state, lake morphology and various
diagenetic processes. Both two-layer oxic/anoxic sediment diagenesis models (e.g. Wang et
al., 2003a, b) as well as more complex one-dimensional, multi-layer, sediment diagenesis
models (e.g. Jørgensen et al., 1982; Boudreau, 1996) have been used to describe the vertical
distribution of nutrients in sediments as well as the fluxes from the sediments to the water
column, which may strongly influence lake water quality. However, due to the complex
nature of these models, they are typically only applied to sediment cores collected from a
single lake (e.g. Van Rees et al., 1996; Schauser et al., 2004).

The main objective of the present study was to quantify the influence of lake trophic state and
morphology, and natural diagenetic processes, on vertical profiles of total phosphorus (TP),
total nitrogen (TN) and total carbon (TC) in sediments of a wide variety of lakes, by

Page 5

5
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
concurrently applying an empirical and a simple, mechanistic diagenesis model. We collected
sediment cores from 14 different lakes (Table 1) in New Zealand and China, ranging from
shallow to deep, and from oligotrophic to highly eutrophic, and derived both an empirical and
a mechanistic model to describe the vertical TP, TN and TC concentration profiles observed
in these cores. Parameter values, obtained by fitting the empirical and mechanistic models to
observed vertical profiles of TP, TN and TC, could then be examined for correlations with
trophic status attributes, lake morphology and a range of diagenetic parameters.

Methods
Sampling sites
Two intact sediment cores were collected from the deep basins in each of the 14 lakes,
targeting areas conforming to accumulation bottoms (c.f. Håkanson and Jansson, 1983). The
two cores were collected at similar depths, but at two different sites within the deep basin of
each lake, in order to capture some of the natural spatial variability within the deep basins.
Twelve of the lakes are deep (zmax >13.5-125m) and lie within the Bay of Plenty Region
(known as the Rotorua lakes), North Island of New Zealand. Two shallow lakes were
selected, including Lake Te Waihora (Lake Ellesmere) in the Canterbury Region, South
Island of New Zealand, and Lake Taihu in the Jaingsu Province, China. Attributes of trophic
status were available for each of the lakes.

Sampling methods
Sampling was conducted in all 14 lakes during the period March 2006 to January 2007. The
sediment cores were collected using a cylindrical gravity or piston corer, which was designed

Page 6

6
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
to leave cores intact. The surface sediment was visually inspected in each core, and if there
was any evidence of disturbance at the sediment-water interface or in the core profile, the core
was discarded and another one taken. Samples of the sediment were collected from each core
at intervals of 2 cm to a vertical depth that varied from 8 to 38 cm, using a custom-made
slicing chamber, which was designed to minimize the exposure of potentially anoxic sediment
to the atmosphere. Each sample was transferred to 50 mL Vulcan™ centrifuge tubes, which
were sealed and placed on ice until return to the laboratory, where pore-waters were
immediately separated by centrifugation at 4000 rpm for 40 min. For Lake Taihu, samples
were collected in air-tight Zip-lock plastic bags, and pore-waters were not separated from the
solid material. The resulting sub-samples of pore-waters and sediment solids from each lake
were stored frozen (-18 ºC) before further analysis.

Analytical methods
Sediment dry weight fraction was determined by weighing solid samples before and after
drying at 105 ºC for 24 h, and also taking into account pore-water mass. Subsequently,
sediment wet weight was determined as the difference between the bulk weight (total weight
including both solids and pore-water) and the sediment dry weight.

Total phosphorus (TP), iron (Fe) and manganese (Mn) content in the dry sediment was
determined after the solids had been ground with mortar and pestle, and approximately 0.5 g
of each sample had been digested with Aqua Regia (3:1 v:v of 1:5 conc. hydrochloric acid
solution and 1:2 conc. nitric acid solution), based on a modified standard procedure (Martin et
al., 1994). Liquid from the digested solid samples and from pore-water samples acidified with

Page 7

7
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
two drops of conc. hydrochloric acid, were then analyzed for TP, Fe and Mn on an ICP-MS
(model ELAN DRC II).

Total carbon (TC) and nitrogen (TN) content in sediment solids was determined by sub-
sampling approximately 0.25 g of the dry sediment solids and analyzing by combustion
(LECO TruSpec model CN Determinator). Most of our study lakes have non-calcareous
sediments with total carbon content closely related to the sediment organic content (McColl,
1977). Concentrations of TC and TN were determined for every second vertical sub-sample in
20 out of the 28 cores; for the remaining cores only the surficial sediments were analyzed.

Total nitrogen in pore-waters was analyzed spectrophotometrically with a Lachat Instruments
flow injection analyzer (model QuikChem 8000 FIA+) following persulphate digestion (Ebina
et al., 1983).

Mathematical and statistical methods
The concentration of TP, TN, TC, Fe and Mn in the dried sediment was calculated for each 2
cm interval in each individual core. The pore-water fraction was excluded from further
statistical analysis as we found this fraction to be negligible relative to the total mass of
elements in the sediments (i.e., pore-waters accounted for an < 1% of the average TP
concentration across all samples collected), and the mechanistic diagenesis model used to
describe vertical profiles is only valid for either the solid or the solute fraction. Most vertical
profiles of the measured elements in the solid sediments showed an appearance of
exponentially decreasing concentrations with depth in the sediments (as demonstrated by TP

Page 8

8
155
156
157
158
159
160
concentration profiles in Fig. 1). We therefore set up an empirical exponential model to
reproduce the vertical distribution of TP, TN, TC, Fe and Mn in (mg kg-1 dry wt), and to
quantify three characteristic profile parameters: the profile surface concentration (β + γ) at the
sediment-water interface, the background concentration (γ) and the vertical decay coefficient
α (cm-1):

)exp()(zzCi
⋅−⋅+=αβγ
(1) 161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177

where Ci(z) is the concentration of TP, TN, TC, Fe or Mn at vertical depth z (cm) in the
sediment (mg kg-1 dry wt). Values of γ, β and α were calculated for each individual sediment
core and for each element (TP, TN, TC, Fe and Mn) using Eq. (1) to fit the observed
geochemical profiles. Goodness of fit of models was tested using Root-Mean-Square-Error
(RMSE) values and Pearson correlation coefficients (r). The RMSE value for each profile
model was minimized by calibrating γ, β and α using Solver in Microsoft Excel, after which
Pearson correlation coefficients were calculated. Solver uses a generalized reduced gradient
non-linear optimization code to minimize model error, thereby searching for and converging
on a minimum in the RMSE value space. In order to evaluate the influence of sediment
compaction on vertical profiles of TP, TN, TC, Fe and Mn, γ, β and α values were also
calculated for wet weight profiles. Each of these three empirical parameters could then be
examined for correlations with trophic status attributes, including water column
concentrations of TP, TN, chlorophyll a (Chl a) and Secchi depth, lake morphology and a
range of diagenetic parameters.

Page 9

9
178
179
180
181
182
183
184
To interpret the three parameters given by the empirical exponential model from a diagenetic
perspective, and to quantify the influence of a range of diagenetic processes on these
parameters, we first consider the general diagenetic advection-diffusion-reaction (ADR)
equation (Berner, 1980; Boudreau, 1997) for the mass balance of solid organic matter (OM),
where burial and biodiffusive mixing are the transport processes, and OM decays with a first
order kinetic rate:

OM
OM
z
∂
OM
2
b
OM
t
∂
Ck
C
w
z
C
∂
D
C
⋅+
∂
⋅−
∂
⋅=
∂
2
(2) 185
186
187
188
189
190
191
192

where COM is the concentration of organic matter in mg kg-1 dry wt, Db is the biodiffusion
coefficient in cm2 yr-1, w is the advective velocity for solids (also referred to as a burial rate
and assumed to be equal to the net sedimentation rate or sediment accumulation rate) in cm
yr-1 and k is a first order kinetic rate coefficient in yr-1 for the breakdown of OM. If we
assume steady state mass-conservation, Eq. (2) simplifies to:


0
2
2
=⋅+
∂
∂
⋅−
∂
∂
⋅
OM
OM
z
OM
b
Ck
C
w
z
C
D
(3) 193
194
195
196
197
198

We can solve Eq. (3) as a second order ordinary differential equation (Boudreau, 1997;
DiToro, 2001; Meysman et al., 2005a), given a constant flux boundary at the sediment-water
interface:

Page 10

0
0
) 1 (
⋅
=
⎥⎦
⎤
⎢⎣
⎡
⋅+
∂
∂
⋅−⋅−=
z
OM
OM
z
b OM
Cw
C
DF
φρ
(4) 199
200
201
202
203
204

where is the constant flux of OM to the sediment-water interface in μg cm-2 yr-1, ρ is the
solid sediment density in g cm-3 and φ the porosity; and a zero-gradient boundary in the deep
sediments:
0
OM
F

0
=
⎥⎦
⎤
⎢⎣
⎡
∂
∂
∞→
z
OM
z
C
(5) 205
206
207
208

The analytical solution to Eq. (3) then becomes:

⎟⎟
⎠
⎟
⎞
⎜⎜
⎝
⎜
⎛
⋅
⋅
⋅⋅+
D
−
⋅
⎥⎦
⎤
⎢⎣
⎡
⋅⋅++
⋅
⋅
−⋅
=
z
kDww
kDww
F
zC
b
b
b
OM
OM
2
4
exp
4
2
)1 (
1
)(
2
2
0
φρ
(6) 209
210
211
212
213
214

Finally, if we assume that OM can be divided into a labile and a refractory fraction, the latter
implicitly also accounting for inorganic matter (where k ~ 0 for refractory OM) the steady
state concentration profile becomes:

⎥
⎦
⎥
⎥
⎥
⎤
⎢
⎣
⎢
⎢
⎢
⎡
⎟⎟
⎠
⎟
⎞
⎜⎜
⎝
⎜
⎛
⋅
⋅
⋅⋅+
D
−
⋅
⎥⎦
⎤
⎢⎣
⎡
⋅⋅++
⋅
+⋅
−⋅
=
z
kDww
kDww
F
w
F
zC
b
b
b
LR
totalOM
2
4
exp
4
2
) 1 (
1
)(
2
2
,
φρ
(7) 215

10

Page 11

11
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239

where FR and FL are a constant flux of refractory and labile OM to the sediment-water
interface, respectively. We can now see that Eq. (7) is equivalent to the empirical expression
in Eq. (1).

To determine how the diagenetic parameters of Eq. (7) influence the vertical concentration
profiles of TP, TN and TC, and how they may be related to environmental variables (water
quality, lake depth, etc.), we applied Eq. (7) to the observed concentration profiles of TP, TN
and TC. When calibrating the diagenetic parameters we assumed constant porosity in each
individual sediment core (φ, estimated to range 0.46-0.92 across sediment cores from all
lakes) and a constant sediment solids density (ρ) of 2.5 g cm-3. To estimate burial rates (w) we
recorded the depth to a tephra layer, which was present at a depth between 7 and 31 cm below
the sediment-water interface in most sediment cores collected from the Rotorua Lakes, New
Zealand. The tephra is comprised of ash and mud which were dispersed over an area > 200
km2 over North Island of New Zealand during the volcanic eruption of Mount Tarawera in
1886 (White et al. 1997). For sediment cores where no tephra was present, we used net
sedimentation data from Trolle et al. (2008) for Rotorua lakes and from Wang et al. (2001) for
Lake Taihu. As no burial rate data were available for Lake Te Waihora, we initially assumed
that this large and shallow eutrophic lake had a burial rate similar to that of Lake Taihu. The
parameters FR and FL were initially fitted by assuming a constant mid-range biodiffusion
coefficient (Db) of 0.5 cm2 yr-1 (Meysman et al., 2003) and a first order kinetic rate coefficient
(k) for moderately labile organic matter of 0.4 yr-1 (Luff et al., 2000). Goodness of fit of
models was again tested using RMSE values and Pearson correlation coefficients. From Eq.
(7) it is evident that the value of FR during initial model calibration (converging on a

Page 12

12
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
minimum of the RMSE value space) will be adjusted relative to the observed background
concentrations. Following the initial calibration step, we performed a second calibration of
parameters FR, FL, Db and k. For the second calibration-step of the diagenetic parameters for
the Lake Te Waihora profiles we also included the burial rate (w).

The three parameters given by the empirical model (Eq. 1) were then examined for
statistically significant linear relationships with the diagenetic parameters given by the
mechanistic model (Eq. 7), and subsequently all parameters from the two models were used to
test for relationships with lake trophic state (represented by biological and chemical water
column attributes) and a selection of morphological variables across all the lakes. The current
trophic state was estimated from annual average (based on years 2005-2006) concentrations
of TP, TN and Chl a and Secchi depths for the water column of each lake from monthly
samples.

Results
Observed geochemical profiles of TP, TN, TC, Fe and Mn
The concentration profiles of TP (Fig. 1), TN, TC, Fe and Mn, derived from each of two
sediment cores collected from within the deep basin of each lake, generally showed a similar
pattern in the deep lakes. The variability between the two cores collected from within the
same basin was, however, typically quite high for the shallower lakes (e.g. Rerewhakaaitu,
Rotoehu, Rotorua, Taihu and Te Waihora, Table 1), even though sampling site depths differed
by <1 m within these lakes. The surficial sediment concentrations, represented by the first
discrete horizontal sediment sample slice (0-2 cm) from each sediment core, ranged between

Page 13

13
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
400 and 4,300 mg P kg-1 dry wt for TP; 1,400 and 19,900 mg N kg-1 dry wt for TN; 7,000 and
136,400 mg C kg-1 dry wt for TC; 7,000 and 57,600 mg Fe kg-1 dry wt for Fe and 140 and
28,800 mg Mn kg-1 dry wt for Mn. For most of these elements the concentration decreased
exponentially with sediment depth, until it became near uniform, typically around 15 cm into
the sediments. The depth at which TP concentrations reached this background level tended to
be deeper into the sediments for some of the deep oligo-mesotrophic lakes, e.g., Lake
Okareka at 15-17 cm compared with deep eutrophic Lake Rotoiti (7-9 cm). The range in
background concentrations was generally smaller than that observed in surface sediment
samples, and was between 200 and 1,300 mg P kg-1 dry wt for TP; 800 and 9,200 mg N kg-1
dry wt for TN; 4,400 and 80,600 mg C kg-1 dry wt for TC; 7,500 and 37,700 mg Fe kg-1 dry
wt for Fe and 180 and 9,100 mg Mn kg-1 dry wt for Mn.

The tephra layers appeared to influence the various geochemical profiles; dry matter content
increased in the tephra, while TN and TC concentrations decreased. The TP, Fe and Mn
concentrations in the tephra layers were, however, generally similar to those found in the
overlying lacustrine sediment.

In examining the concentration profiles of TP (Fig. 1), there was no clear separation between
oligotrophic and eutrophic lakes. For example, the TP profiles in sediment cores collected
from deep, oligotrophic Lake Okataina, where anthropogenic impacts are still negligible,
showed a similar vertical distribution to TP profiles from sediments in deep, eutrophic Lake
Rotoiti, which has undergone a period of severe eutrophication during the 1970s (Vincent et
al., 1984) and has since remained eutrophic (Hamilton et al., 2006).

Page 14

14
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308

Model predictions of sediment geochemical profiles
The empirical model used to describe the vertical decay of TP, TN, TC, Fe, Mn and wet
weight with sediment depth, as well as the diagenetic model for profiles of TP, TN and TC,
generally produced close fits to observed geochemical profiles across the 14 lakes (Table 2).
Visual comparison of a subset of the sediment geochemical profiles in cores from lakes
representing eutrophic, mesotrophic and oligotrophic states (Fig. 2) showed irregular
concentration profiles of various elements (e.g., oligotrophic Lake Tarawera; Fig. 2). In
sediment cores that extended into the Tarawera tephra, the tephra values were omitted from
the model fit. As the empirical and the mechanistic models were both based on an exponential
decrease of elemental concentrations with sediment depth, these models explained the same
relative amount of variability in the concentration profiles of TP, TN and TC (Table 2). The
models explained between 51 and 100% of the variability in the vertical profiles of TP, TN
and TC across the 14 lakes. Concentration profiles which were relatively uniformly or linearly
distributed through the sediments (mostly for Fe and Mn) were also reproduced satisfactorily
by the empirical model, with the vertical decay coefficient (α) for these profiles equal to zero
(Table 2).

Geochemical profiles related to diagenetic parameters
We found a strong and significant correlation between the sediment background
concentrations (γ) and the flux of refractory matter (FR) of TP, TN and TC (Table 3), whereas
the surface concentrations (γ + β) of TP, TN and TC were more closely related to the flux of
labile matter (FL). The vertical decay coefficients (α) for TP, TN and TC profiles were most

Page 15

15
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
strongly related to the biodiffusion coefficients (Db), while the vertical decay coefficients for
TN and TC profiles, but not TP profiles, were also significantly and inversely correlated with
the burial rates (w). The vertical decay coefficients for TP were also significantly, though
weakly, related to the vertical decay coefficients for Fe (r = 0.47, p < 0.05, n = 28), Mn (r =
0.46, p < 0.05, n = 28) and wet weight (r = 0.50, p < 0.01, n = 28). However, the correlation
between the vertical decay coefficients for TP and the Fe and Mn profiles cannot be justified
as causation, as the vertical decay coefficients for the Fe and Mn profiles were also strongly
inter-correlated with the vertical decay coefficients for wet weight profiles (p < 0.001).
Vertical decay coefficients for wet weight profiles were also strongly correlated with the
vertical decay coefficients for TN and TC profiles. The first order kinetic rate coefficients (k)
were not significantly correlated with any of the parameters given by the empirical model for
TN and TC profiles, but were significantly related to both the surface concentrations and the
vertical decay coefficients for the TP profiles (Table 3).

Geochemical profiles related to morphological, chemical and biological variables
We generally found no, or only weak, correlation between the parameters given by the two
models and lake water quality data, represented by annual mean TP, TN, Chl a concentrations
and Secchi depth (Table 3). The biodiffusion coefficients and vertical decay coefficients of
TN profiles were, for example, significantly correlated with water column TN concentrations.
The coring site depth, lake mean depth, surface area and catchment area were generally not
significantly correlated with any of the parameters given by the two models (p > 0.05).
However, the coring site depth to lake mean depth ratio was significantly correlated with the
burial rates and the flux of refractory TC (Table 3), and also inversely correlated with the
vertical decay coefficients for wet weight profiles. The vertical decay coefficients for wet