wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Controlling harmful cyanobacterial blooms ina hyper-eutrophic lake (Lake Taihu, China): The needfor a dual nutrient (N & P) management strategy

Hans W. Paerl a,*, Hai Xu b, Mark J. McCarthy c,d, Guangwei Zhu b, Boqiang Qin b,Yiping Li e, Wayne S. Gardner d

aUniversity of North Carolina at Chapel Hill, Institute of Marine Sciences, 3431 Arendell Street, Morehead City, NC 28557, USAb State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography & Limnology, Chinese Academy of Sciences,

73 East Beijing Road, Nanjing 210008, PR ChinacUniversite du Quebec a Montreal, Departement des sciences biologiques, Montreal, Quebec H3C 3P8, CanadadUniversity of Texas at Austin, Marine Science Institute, 750 Channel View Drive, Port Aransas, TX 78373, USAeDepartment of Environmental Science and Engineering, Hohai University, 1 Xikan Road, Nanjing 210098, PR China

a r t i c l e i n f o

Article history:

Received 16 February 2010

Received in revised form

26 August 2010

Accepted 14 September 2010

Available online 29 September 2010

Keywords:

Eutrophication

Nitrogen

Phosphorus

Cyanobacteria

China

Nutrient management

* Corresponding author. Tel.: þ1 252 726 684E-mail address: hpaerl@email.unc.edu (H

0043-1354/$ e see front matter ª 2010 Elsevdoi:10.1016/j.watres.2010.09.018

a b s t r a c t

Harmful cyanobacterial blooms, reflecting advanced eutrophication, are spreading globally

and threaten the sustainability of freshwater ecosystems. Increasingly, non-nitrogen (N2)-

fixing cyanobacteria (e.g.,Microcystis) dominate such blooms, indicating that both excessive

nitrogen (N) and phosphorus (P) loads may be responsible for their proliferation. Tradi-

tionally, watershed nutrient management efforts to control these blooms have focused on

reducing P inputs. However, N loading has increased dramatically in many watersheds,

promoting blooms of non-N2 fixers, and altering lake nutrient budgets and cycling char-

acteristics. We examined this proliferating water quality problem in Lake Taihu, China’s

3rd largest freshwater lake. This shallow, hyper-eutrophic lake has changed from bloom-

free to bloom-plagued conditions over the past 3 decades. Toxic Microcystis spp. blooms

threaten the use of the lake for drinking water, fisheries and recreational purposes.

Nutrient addition bioassays indicated that the lake shifts from P limitation in wintere

spring to N limitation in cyanobacteria-dominated summer and fall months. Combined N

and P additions led to maximum stimulation of growth. Despite summer N limitation and P

availability, non-N2 fixing blooms prevailed. Nitrogen cycling studies, combined with N

input estimates, indicate that Microcystis thrives on both newly supplied and previously-

loaded N sources to maintain its dominance. Denitrification did not relieve the lake of

excessive N inputs. Results point to the need to reduce both N and P inputs for long-term

eutrophication and cyanobacterial bloom control in this hyper-eutrophic system.

ª 2010 Elsevier Ltd. All rights reserved.

1. Introduction eutrophication. These blooms are increasing worldwide and

Harmful (toxic, food web-disrupting) cyanobacterial blooms

(CyanoHABs) are a troubling indicator of advanced

1; fax: þ1 252 726 2426..W. Paerl).ier Ltd. All rights reserved

represent a serious threat to drinking water supplies, and the

ecological and economic sustainability of our largest fresh-

water ecosystems (Reynolds, 1987; Paerl, 1988; Carmichael,

.

Fig. 1 e Upper: Satellite (MODIS) image of Lake Taihu, the

nearby cities of Wuxi (N), Souzhou (E), and Shanghai (E),

near the mouth of the Yangtze River on 7 June, 2007, when

a massive cyanobacterial (Microcystis spp.) bloom that

covered almost the entire lake (image courtesy NASA).

Lower: Photograph of a Microcystis bloom on the north side

of the lake on 24 October, 2009 (photograph by H. Paerl).

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 31974

2001; Huisman et al., 2005). Examples include Lake Erie (North

America), LakeWinnipeg, Canada, Lake Victoria, the largest of

the African rift lakes, Lakes Biwa and Kasimagaura, Japan’s

largest lakes, and Lake Taihu, the 3rd largest freshwater lake

in China. Anthropogenic nutrient over-enrichment of these

and smaller lakes and reservoirs has been linked to CyanoHAB

proliferation (Paerl et al., 2001; Huisman et al., 2005). Nitrogen

(N) and phosphorus (P) are the key nutrients of concern

(Likens, 1972; Schindler, 1977; Paerl, 2008). Inputs of both

nutrients to natural waters have increased dramatically in the

post-World War II era. A central question facing water

researchers and managers is; which nutrient(s) control Cya-

noHAB production and proliferation in impacted systems?

The answer to this question is of immense ecological and

economic importance, because it dictates the strategies and

costs involved in mitigating this serious water quality

problem.

Phosphorus has been implicated traditionally as having

a central role in the control of freshwater primary production

(Likens, 1972) and CyanoHAB bloom formation (Paerl, 1988,

2008). This conclusion is particularly relevant to nitrogen

(N2) fixing CyanoHABs, since they may satisfy their own N

requirements (Smith, 1983, 1990). As a result, P input restric-

tions have been implemented widely since the 1960s. These

reductions have slowed eutrophication rates and reduced

algal bloom potentials (Schindler, 1977). However, nutrient

loading dynamics have changed considerably since the 1960s.

Agricultural, urban and industrial nutrient sources have

accelerated rapidly (Vitousek et al., 1997; Galloway and

Cowling, 2002). These sources are treated more effectively

for removal of P than N before being discharged, leading to

higher N than P loading to already nutrient-stressed water

bodies (Paerl, 1997; Rabalais, 2002; Boyer et al., 2004). Excessive

N loads have promoted non-N2 fixing CyanoHABs, including

expanding blooms of the toxin producing, colonial, surface-

dwelling (buoyant) cyanoHABMicrocystis, and the filamentous

genera Lyngbya (non-N2 fixing strains) and Oscillatoria (Paerl,

2008, 2009).

A severely impacted large lake system is Taihu (meaning

“great lake” in Mandarin), with an area of 2338 km2 and

a volume of 4.4 billion m3 (Pu and Yan, 1998; Qin et al., 2007,

2010). This shallow (mean depth 1.9 m) polymictic lake is

located in the Yangtze River delta; themost rapidly developing

region in China (Fig. 1). Approximately 40 million people live

within the Taihu watershed. The Taihu Basin accounts for

only 0.4% of China’s land area, but the region accounts for 11%

of its Gross Domestic Product (Qin et al., 2007). The lake is

a key drinking water, fishing and tourism resource for the

region. However, it also serves as a waste repository for urban,

agricultural and industrial segments of the local economy

(Guo, 2007; Qin et al., 2007). Consequently, Taihu has experi-

enced accelerating eutrophication over the past 3 decades

(Qin et al., 2007, 2010). During this period, it has changed from

a mesotrophic, diatom-dominated lake to hyper-eutrophic,

cyanobacteria-dominated system, with Microcystis blooms

now occurring regularly throughout much of the lake (Chen

et al., 2003a,b) (Fig. 1). These blooms have caused environ-

mental, economic and societal problems, including a threat to

potable water supplies for approximately 10 million

consumers (Guo, 2007).

Microcystis and other non-N-fixing cyanobacteria are

effective competitors for reduced N forms, especially ammo-

nium (NH4þ) (Kappers, 1980; Blomqvist et al., 1994), which is

rapidly regenerated in the water column and sediments of

these shallow systems. Cyanobacteria capable of N2 fixation

can also assimilate NH4þ when available. For example, Shel-

burne Pond in Vermont was dominated by N2 fixers, but

characterized by low rates of N fixation (<9% of total N uptake)

and NOx uptake (<5%), but high acquisition rates of NH4þ

(82e98%) in summer (Ferber et al., 2004). Thus, the presence of

N2 fixers does not ensure that N2 fixation is a significant N

source; regenerated NH4þ may be a key N source for sustaining

blooms, even among N2 fixing groups.

Although Microcystis blooms dominate, heterocystous,

filamentous genera capable of N2 fixation (Anabaena, Aphani-

zomenon) are also present in the water column and as over-

wintering cells in the sediments of Taihu. The N2 fixing genera

increase in numeric importance toward the lake center, i.e.,

away from shoreline locations, although they remain sub-

dominant toMicrocystis (Chen et al., 2003a,b).While N2 fixation

rates have not been measured in the lake, these observations

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3 1975

suggest that N2 fixing conditions may increase from the lake’s

shorelines to its open waters, as hypothesized by McCarthy

et al. (2007). These observations also suggest that nutrient

availability and limitations may vary spatially in the lake.

Temporal differences may also be important because water-

shed N and P inputs exhibit seasonal patterns (Qin et al., 2007,

2010; Xu et al., 2010).

The fact that both N2 fixing and non-N2 fixing cyanoHABs

coexist in hyper-eutrophic Taihu has raised questions as to

whether N, P or both N and P inputs should be reduced to

control blooms. Based on research in other lakes, primary

production and bloom formation may be controlled by both N

and P inputs, either contemporaneously or sequentially, with

different individual nutrients being limiting at different times

of the year (Dodds et al., 1989; Elser et al., 2007; Havens et al.,

2001; Kronvang et al., 2005; Jeppesen et al., 2007; North et al.,

2007; Lewis and Wurtsbaugh, 2008; Ozkan et al., 2009; Xu

et al., 2010). These findings have been challenged recently,

based on the premise that lake eutrophication cannot be

controlled by reducing N inputs (Schindler et al., 2008). This

conclusion relies on the observation that many systems

exhibiting advanced eutrophication also contain significant

N2 fixing cyanoHAB populations, and the assumption that N

fixed by these populations can meet ecosystem N require-

ments. However, diverse studies show that only a fraction,

usually far less than 50% of ecosystem-level N demands, is

met by N2 fixation (Howarth et al., 1988; Paerl, 1990; Lewis and

Wurtsbaugh, 2008). In fact, a recent re-examination of

Schindler et al.’s (2008) data on P-fertilized Lake 227 indicates

that N2 fixation does not meet ecosystem N demands (Scott

and McCarthy, 2010). In other cases, N2 fixation rates are

very low and supply little new N, even when N fixing cyano-

bacteria are dominant (Ferber et al., 2004). Furthermore,

factors in addition to stoichiometric N:P ratios (Smith, 1983,

1990) control this energy-demanding process in aquatic

ecosystems (Howarth et al., 1988; Paerl, 1990; Forbes et al.,

2008). Reversing eutrophication and reducing CyanoHABs in

a range of lakes have required either reduction of only P

(Schindler, 1977) or both N and P inputs (Kronvang et al., 2005;

Jeppesen et al., 2007). Hence, nutrient reduction strategies

appear system-specific.

We examined effects of individual and combined N and P

additions on phytoplankton growth (based on chlorophyll

a concentrations) in Taihu, using short-term (up to 6 days)

nutrient addition bioassays incubated under natural light and

temperature conditions. These bioassays provide a rapid

assessment of nutrient limitation characteristics, i.e., imme-

diate growth responses, rather than predicting long-term

phytoplankton succession patterns (Paerl and Bowles, 1987;

Piehler et al., 2009). We also evaluated NH4þ regeneration and

potential uptake rates within the context of dominance by

non-N2 fixing cyanoHABs.

2. Methods and materials

2.1. Location and field sites

Lake Taihu is located approximately 150 kmwest of Shanghai;

with the lake center coordinates at 31�100000N, 120�90000E. The

Taihu drainage basin is 36,500 km2 (Fig. 2). The lake has more

than 30 input sources, ranging from rivers to small streams

andman-made drainage canals. Water exits the southeastern

corner of Lake Taihu via the Taipu River, which drains through

Shanghai into the East China Sea (Figs. 1 and 2).

Field monitoring and bioassay water collection sites were

located in one of the northern bays, Meiliang Bay (Inner Bay,

Outer Bay), and the lake proper (Main Lake) (Fig. 2). Other lake

sites sampled included the least bloom-impacted eastern

regions of the lake (ELT stations) (Fig. 2). Meiliang Bay was

chosen as a focal point because it is the site of recurring and

intensifying Microcystis spp. blooms (Chen et al., 2003a,b; Qin

et al., 2007). Meiliang Bay receives freshwater inputs from

the Liangxi and Zhihu Gang rivers, which drain untreated

wastewater from factories, residential and agricultural areas.

The named rivers in Fig. 2 contribute more than 85% of the

lake’s freshwater inflow.

2.2. Nutrient inputs to Lake Taihu

Data were obtained from the Taihu Basin Authority, Ministry

of Water Resources, China, for 30 primary tributaries from

2000 to 2005, includingmonthly flow discharge and TN and TP

concentrations, to estimate N and P loads discharged to the

lake (Fig. 2). These tributaries with 30 monitoring stations

accounted for approximately 85% of the total runoff input to

the lake (Qin et al., 2007). TN or TP load for each month was

calculated by the following formula:

Fi ¼ 2:592$X30

j¼1

Cij$Qij (1)

where Fi is TN or TP load for ithmonth (i¼ 1e12) inmetric tons,

Cijis the TN or TP concentration for ithmonth and jth tributary

( j¼1e30) inmg/L.Qij is theflowdischarge for ithmonthand jth

tributary ( j¼1e30) inm3/s. 2.592 is the factor for convertingg/s

to ton/mo. Spring, summer, autumn and winter in this paper

represent MarcheMay, JuneeAugust, SeptembereNovember,

and DecembereFebruary, respectively.

We estimated historic changes in atmospheric N loads to

the lake based on data derived from regional atmospheric

deposition models (Lelieveld and Dentener, 2000) and direct

measurements (Zhai et al., 2009).

2.3. Lake environmental measurements

Monthly samples for nutrients, chlorophyll a, and phyto-

plankton identification and enumeration were collected at the

Meiliang Bay and Main Lake locations (Fig. 2). Water samples

for nutrient addition bioassays were collected at the Inner Bay

location. For logistical reasons, the bioassays were incubated

at a lake location near the Taihu Laboratory for Lake

Ecosystem Research (TLLER), Nanjing Institute of Geography

and Limnology, Chinese Academy of Sciences (Fig. 2).

Monthly physical, chemical, and biological parameters,

including surface water temperature, dissolved oxygen, pH,

and electrical conductivity, were measured using a Yellow

Springs Instruments (YSI) 6600multi-sensor sonde. Integrated

water sampleswere taken using a 2-m long, 0.1-mwide plastic

tube with a 1-way valve.

Fig. 2 e Map of Lake Taihu. The lake’s largest tributaries are named. The Taipu River is the main river discharging water

from the lake. The Wangyu River periodically delivers Yangtze River water to the lake (see Fig. 1 for the location of the

Yangtze River relative to Taihu). Sampling locations are indicated, as is the Taihu Laboratory for Lake Ecosystem Research

(TLLER). Insert shows the location of Taihu in the PR China.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 31976

2.4. Nutrient analyses

Water samples were analyzed for total nitrogen (TN), total

dissolved nitrogen (TDN), dissolved inorganic nitrogen (DIN;

ammonium (NH4þ) þ nitrate (NO3

�) þ nitrite (NO2�)), total phos-

phorus (TP), total dissolved phosphorus (TDP), and dissolved

inorganic phosphorus (DIP). DIP was determined using the

molybdenum bluemethod (APHA, 1995). NH4þ was determined

using the indophenol bluemethod, andNO3� andNO2

�with the

cadmium reduction method (APHA, 1995). TP, TDP, TN, and

TDNweredeterminedusing a combined persulphate digestion

(Ebina et al., 1983), followed by spectrophotometric analysis as

for DIP and NO3�. Particulate nitrogen (PN) was obtained by

subtracting TDN from TN, and particulate phosphorus (PP) by

subtracting TDP fromTP. Analytical errorswere determined as

the average % coefficients of variation of triplicates. Average

errors for PP and PN were 6.3% and 5% respectively.

2.5. Biological measurements

Phytoplankton samples were preserved with Lugol’s iodine

solution (2% final conc.) and settled for 48 h. Phytoplankton

species were identified and counted according to Hu et al.

(1980). Chlorophyll a (Chl a) concentrations were determined

spectrophotometrically, following extraction in 90% hot

ethanol (Papista et al., 2002).

2.6. Nutrient limitation bioassay experiments

In situ nutrient addition bioassays were performed seasonally

in 2008e2009 on “inner bay” water (Fig. 2) to examine nutrient

limitation of the natural phytoplankton community. Water

samples were collected from 0.2 m below the surface using

0.01 N HCl-washed and then lake water-rinsed 20 L poly-

ethylene carboys. Water was screened through 200 mm mesh

to remove large zooplankton and dispensed into acid (0.01 HCl)

and then lake water washed 1-L polyethylene Cubitainers

(HedwinCo.). Cubitainers are chemically inert, unbreakableand

transparent (80% PAR transmittance). The methodology and

deployment for Cubitainer bioassays followed Paerl and Bowles

(1987). At the start of each experiment (To), water sampleswere

analyzed for Chl a and nutrients. Three treatments were con-

ducted, in addition to a control (no nutrient additions): (1) N

addition (þN) (2) P addition (þP) (3) N and P addition (þNP).

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3 1977

Nitrogen was added as KNO3, reflecting the dominant form of

inorganic N in Taihu. NH4Cl was also used as an N source in

summer 2009 to compare phytoplankton growth responses to

different N forms. Phosphoruswas added as K2HPO4$3H2O. The

final concentration of N was 2.00mg N/L, and the final concen-

tration of P was 0.50 mg P/L. These concentrations approxi-

mated the values of these nutrient forms in the lake during

maximumdischarge periods (Pu andYan, 1998; Qin et al., 2007).

All treatments were conducted in triplicate. Following

nutrient additions, Cubitainers were incubated in situ near

the surface for 4 or 6 days by placing them in a floating

frame suspended off a pier at TLLER (Fig. 2). This approach

provided natural light, temperature and surface wave action

conditions. One layer of neutral density screening was

placed over the frame to prevent photoinhibition during the

incubations. Following deployment, Cubitainers were sub-

sampled at 2e3 d intervals for Chl a and nutrient

concentrations.

2.7. Water column NH4þ regeneration and potential

uptake rates

Water column NH4þ regeneration and potential uptake rates

in Taihu, as described in McCarthy et al. (2007), were con-

ducted to provide insights into consequences of unmitigated

N discharges into Taihu. Previously published water column

N cycling rates were determined in late summer 2002 from

four sites in northern Taihu (McCarthy et al., 2007). The N

and P addition bioassays described in the present study

were conducted on water collected at sites corresponding to

“inner bay” and “main lake” in McCarthy et al. (2007). Water

column NH4þ recycling rates also were determined at the

main lake station and four sites in East Lake Taihu (ELT) in

January 2004 and all northern Taihu and ELT locations in

May 2004 (Fig. 2). The ELT sites were located in the south-

eastern portion of the lake, near the outflow of the lake into

the Taipu River. Methodological details are provided in

McCarthy et al. (2007).

Briefly, water samples were amended with saturating

levels of 15NH4þ and incubated at ambient temperature and

light for w24 h. Total NH4þ concentration and isotope ratios

were determined using high performance liquid chromatog-

raphy (HPLC; Gardner et al., 1995), and regeneration and

potential uptake rates were calculated using the isotope

dilution technique (Blackburn, 1979; Caperon et al., 1979).

Since Lake Taihu is shallow and well-mixed, rates were

extrapolated to the whole water column to estimate a total

internal N load from water column recycling processes.

Fig. 3 e Upper: Seasonal total nitrogen (TN) loading to Lake

Taihu from its tributaries, covering a period from 2000 to

2005, inmetric tons of N. Lower: Seasonal total phosphorus

(TP) loading, in metric tons of P, to Lake Taihu from its

tributaries, from 2000 to 2005.

2.8. Statistical analyses

Differences in the growth responses between various treat-

ments were analyzed by one-way ANOVA. Post Hoc Multiple

Comparisons of treatment means were performed by Tukey’s

least significant difference procedure. Untransformed data in

all cases satisfied assumptions of normality and homosce-

dasticity. Statistical analysis was performed using the SPSS

13.0 statistical package for personal computers, and the level

of significance used was at p < 0.05 for all tests.

3. Results and discussion

3.1. Nutrient inputs and concentrations

Seasonal and annual surface water TN and TP loads to Taihu

during 2000e2005 indicate that the lake received high

amounts of N and P on a year-round basis, with N loads being

higher, relative to P, in the winter (Fig. 3). Ratios (by weight) of

TN to TP loading range from w22 to over 29, indicating rela-

tively N enriched conditions on an annual basis. Lowest N:P

loading ratios occur in summertime, whereas highest ratios

occur in winterespring.

The Taihu basin has, over the past 3 decades, experienced

dramatic increases in population and urbanization (Qin et al.,

2007; Guo, 2007). The high N:P loading reflects the combined

effects of population growth, changes in land use, including

increasing agricultural, and urban and rural wastewater

discharge. These changes and the emphasis of P over N

control in wastewater treatment (Wang andWang, 2009) have

led to increases in N:P loading from cities. In addition,

reflecting a worldwide pattern (Galloway and Cowling, 2002),

chemical N fertilizer use has increased in agricultural regions

0.00

0.05

0.10

0.15

0.20

0.25

0.30

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

Year

TP

(mg/

L)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

TN

(m

g/L

)

TP TN

Fig. 4 e Annual average TN and TP concentrations, as mg/L

of each element, in Lake Tahu. Data are from the Main Lake

station.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 31978

of the Taihu watershed, leading to N relative to P enrichment.

Lastly, atmospheric deposition, which is enriched in N relative

to P (Zhai et al., 2009), is an increasing source of N, accounting

for approximately 30% of the external N loading to the lake.

Atmospheric N inputs are estimated to have undergone a 25-

fold increase over the past 150 years, and this trend has

accelerated during the past 3 decades (Lelieveld and Dentener,

2000) (Supplementary material 1: S-1).

Recent increases in TN and TP loading can be seen as

trends in TN and TP concentrations in the lake (mid-lake

location) (Fig. 4). Increases were evident between the mid

1980s and 2000, a period of rapid population growth and

urbanization in the Taihu watershed (Qin et al., 2007; Guo,

2007). They may have leveled off since the early 2000,

Fig. 5 e Patterns of TN, DIN, TP and DIP, as mg/L of each elemen

possibly a result of persistent droughts (Qin et al., 2007, 2010),

diversion of sewage from the cities of Wuxi and Suzhou to

waterways draining to the East China Sea, and usage of the

Wangyu River canal, which exchanges water between Taihu

and the Yangtze River. The decreases TN loading may have

resulted from a reduction in pollutant input from the water-

shed in 1998. However, at the end of 1998, a pollutant emission

reduction movement (the so called “zero point action”)

launched by the government throughout the entire water-

shed, did not persist over time.

Both TN and TP, as well as DIN and DIP concentrations,

showed strong seasonal variation in Taihu. Maximum TN and

DIN values occurred in winter and spring, whereas minimum

values were observed in summer and autumn during the 2-yr

period (Fig. 5). In contrast, TP and DIP values showed an inverse

pattern,withwinterhaving lowvaluesandsummerhighvalues.

DIN patterns closely tracked TN over time (Fig. 5). Overall, TN

values were higher in Meiliang Bay than in the lake proper,

reflecting large external loads and elevated internal loading.

3.2. Nutrient addition bioassays

The in situ nutrient addition bioassays showed stimulation of

algal biomass production (as Chl a) in response to individual

and combined N and P additions, indicating that nutrient

enrichment enhanced algal growth and bloom potentials.

Nutrient limitation showed strong repeated seasonal patterns

in 2008 and 2009, although different degrees of algal biomass

stimulation were observed between the two years (Fig. 6). P

limitation prevailed in winter and spring, while N limitation

occurred in summer and fall. In most instances, algal growth

responses to N and P additions exceeded those observed with

t, at Inner Bay and Main Lake locations during 2008e2009.

Fig. 6 e Phytoplankton biomass (chlorophyll a) responses

in bioassays conducted in May, July, October, and

December 2008 and May, July and October 2009. Water

samples for bioassays were collected from the surface at

the Inner Bay location in Meiliang Bay. Initial chlorophyll

a content is shown. Responses were for 3-day incubations

in spring, summer, and fall, 6-day incubations in winter

2008, and 2-day incubations in spring, summer, and fall

2009. Mean values are shown. Error bars represent ±1SD of

triplicate samples. Differences between treatments are

shown based on ANOVA post hoc tests (a > b > c;

p < 0.05).

Fig. 7 e Phytoplankton biomass (chlorophyll a) responses

to various nitrogen sources, each added at 2.00 mg N/L, in

bioassays conducted in August 2009, using Inner Bay

water from Meiliang Bay. Response is 2-day chlorophyll

a average. Error bars represent ±1SD of triplicate samples.

Differences between treatments are shown based on

ANOVA post hoc tests (a > b > c; p < 0.05).

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3 1979

N and P alone (Fig. 6). This “synergistic” N and P effect was

most pronounced in summer and fall, when algal growth

rates, and hence nutrient demands, were highest.

Nutrient limitation patterns followed inverse concentra-

tion patterns of dissolved inorganic N (DIN) and P (DIP). DIN

concentrations were high during winter and spring, reflecting

high N inputs, while DIP concentrations were at their lowest

levels (Fig. 5). As a result, DIP proved to be the limiting nutrient

at this time. Then, DIN decreased rapidly as growth and bloom

conditions improved in late spring and summer with

increased light levels and temperatures. In contrast, DIP

concentrations remained quite high and actually increased

during summer bloom periods. The summer DIP increases

may relate to elevated pH conditions (due to photosynthetic

CO2 demand by blooms), which can enhance DIP release from

the sediments (Andersen, 1975; Xu et al., 2010). Further,

Microcystis can store P during sedimentary phases and

assimilate N in larger proportion to P during bloom phases

(Ahn et al., 2002). This uneven nutrient assimilation pattern

can lead to decreases in N:P and enhance N limitation. It

appears that DIN availability during this period determines

the magnitudes and duration of booms.

The availability of iron (Fe), a nutrient required for photo-

synthetic growth and N2 fixation, may have played an addi-

tional role in controlling phytoplankton production. However,

when Fe (as EDTA-chelated and non-chelated Feþ2) was added

alone and in combination with N and P to Inner Bay and open

lake water samples during summer 2009, no significant

stimulatory (or inhibitory) effects of Fe were observed (not

shown).

Parallel microscopic determinations of phytoplankton

biomass agreed with chlorophyll a results (Xu et al., 2010),

confirming that Chl a was a good indicator of phytoplankton

biomass response in the bioassays. Additional total particu-

late organic carbon measurements made on the bioassays

confirmed that Chl a responses reflected true increases in

phytoplankton biomass.

Microcystis spp. remained the dominant bloom-forming

cyanobacteria during the summer-fall bloom periods of both

years, despite chronic N limitation. These N limited periods

should have provided optimal conditions for N2 fixing genera

(i.e., Anabaena, Aphanizomenon) to become dominant (Smith,

1983, 1990; Schindler et al., 2008), but this situation did not

develop, even though DIP remained plentiful (Figs. 4 and 5).

Possible explanations for this result include; (1) superior ability

of Microcystis spp. to compete for NH4þ and P from sediments

(Kappers,1980)andwatercolumnregeneration (Blomqvistetal.,

1994), and (2) mutually-beneficial bacterialecyanobacterial

interactions in the “phycosphere” of Microcystis spp. colonies,

which can enhance nutrient cycling and growth of “host”

Microcystis populations (Paerl and Pinckney, 1996).

Bioassays illustrated thatMicrocystis spp. competed for DIN

effectively, especially for NH4þ. Per amount of N added,

ammonium stimulated significantly more algal biomass

formation than nitrate (Fig. 7). The extent to which natural

Microcystis populations dominated the absolute uptake of NH4þ

was not determined, but phytoplankton biomass was domi-

nated (>80%) by Microcystis in bioassays. This mechanism

helps explain the persistence of Microcystis during periods of

low DIN concentrations. In addition, Microcystis’ ability to

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 31980

adjust its vertical position by buoyancy compensation

(Reynolds, 1987) may enable exploitation of the entire water

column, taking advantage of regenerated as well as exter-

nally-supplied (atmospheric, surface runoff) N sources.

The results indicate that inputs of both nutrients should be

reduced to control bloom formation and magnitude. Algal

biomass production may be controlled by P availability in the

spring, while N availability may determine the magnitude,

spatial extent and duration of the bloom during summer-fall

when the bloom potential is highest. Nutrient co-limitation

was observed during all periods; i.e., combined enrichment

with N and P led to higher magnitudes of biomass formation

than either N or P alone. This result suggests that N and P

supplies are closely balanced with regard to the requirements

for supporting and promoting eutrophication and bloom

formation.

3.3. Water column NH4þ regeneration and potential

uptake

Water column NH4þ regeneration and potential uptake rates

showed seasonal and spatial differences (S-1). In January, light

uptake rates were significantly lower ( p < 0.01; ANOVA) than

those in September (main lake) and May (main lake and ELT

sites). In northern Taihu, only the inner bay had a seasonal

difference in regeneration rates, although the difference was

large (0.89 mmol N/L h in May versus 0.19 mmol N/L h in

September). In ELT, uptake and regeneration rates were lower

than the Meiliang Bay sites, but this pattern was expected

since ELT is dominated by submerged aquatic vegetation

(SAV) rather than phytoplankton.

To scale regeneration rates and to compare with external

loads, sites were first regrouped based on location within the

lake. Lake Taihu has a surface area of 2338 km2, but Meiliang

Bay and ELT account for only 100 and 131 km2, respectively, of

the total surface area. The outer bay site was located at the

interface between Meiliang Bay and the main lake and is

15e20 km from river discharges. Thus, this site was grouped

with the main lake station. After regrouping the sampling

sites, volumetric regeneration rates (see S-1) were converted

to areal rates using water depth and extrapolated based on

surface area of the appropriate lake region.

Extrapolatedwater columnNH4þ regeneration rates suggest

that 3.77� 107 kg N/yr are regenerated as NH4þ in Meiliang Bay,

where the most severe Microcystis blooms occur. Despite

actual regeneration rates being an order ofmagnitude lower in

the main lake than Meiliang Bay, the large surface area of the

central basin results in 6.57 � 107 kg N/yr regenerated as NH4þ.

As expected, ELT plays a minor role in total N recycling in the

water column (0.13 � 107 kg N/yr). The sum of these annual

regeneration estimates is about 400% of total estimated N

loading to the lake (2.5 � 107 kg N/yr; James et al., 2009).

However, these estimates include only N regenerated as NH4þ.

No estimates of the NH4þ proportion of the total N load are

known, but the proportion is presumably small relative to

oxidized (i.e., NO3�) and organic N forms. Internal N cycling is

important for the maintenance and species succession of

cyanobacteria blooms in Taihu, especially as it pertains to

Microcystis spp. (McCarthy et al., 2007). Atmospheric deposi-

tion is a significant additional source of bioavailable N in the

lake (Zhai et al., 2009) (S-2). For example, NH4þ and NO3

-con-

centrations of a rainwater sample collected in May 2004 were

370 and 146 mM, respectively (McCarthy and Gardner, unpub-

lished data).

Direct denitrification measurements in Meiliang Bay and

the main lake in late summer 2002 (McCarthy et al., 2007)

and Meiliang Bay, the main lake, and ELT from January to

May 2004 (McCarthy and Gardner, unpublished data) were

extrapolated to estimate a lake-wide denitrification rate.

This rate ranged from 7360 kg N/km2 per year when esti-

mated by net N2 flux to 26,700 kg N/km2 per year when

estimated using 15NO3� addition assays. Both estimates were

obtained from continuous-flow incubations of intact sedi-

ment cores. Rates from 15NO3� additions should be qualified

as potential rates, whereas rates from net N2 flux would

include any N2 fixing activities, which were not significant in

sediments of this lake (McCarthy et al., 2007). Therefore, net

N2 flux represents the best estimate of denitrification and

accounts for 66.2% of external N loading. This N loss via

denitrification would not account for the N recycled in the

water column. In late summer, Meiliang Bay and main lake

sediments are an N source to the water column (McCarthy

et al., 2007). Sediments also are an N source in ELT and the

main lake in January and May (McCarthy and Gardner,

unpublished data). However, sediments in Meiliang Bay were

a strong N sink in January and May. These patterns suggest

that late summer cyanobacteria blooms rely, in part, on

nutrients released from sediments (McCarthy et al., 2007).

Depth averaged water column NH4þ regeneration rates for the

shallow water column imply that water column regeneration

supplies a greater amount of N (5-fold more on an areal

basis) than sediments for cyanobacterial assimilation in the

summer (McCarthy et al., 2007).

In addition to the importance that total N loads play in

determining rates of eutrophication, the supply rates and

ratios of various N forms help structure microalgal commu-

nities mediating freshwater primary production (Paerl, 1988;

McCarthy et al., 2007, 2009). For example, the ratio of NH4þ to

oxidized N was related to the proportion of cyanobacteria

comprising the total phytoplankton community of Lake

Okeechobee, FL, USA (McCarthy et al., 2009). While non-N2

fixing cyanobacteria, such as Microcystis, compete effectively

for reduced N (Blomqvist et al., 1994), N2 fixing cyanobacteria

also assimilate ammonium preferentially if it is available

(Ferber et al., 2004). Ammonium and other reduced N forms,

such as dissolved free amino acids, are more available than

oxidized N forms (nitrate and nitrite) to bacteria (Vallino et al.,

1996) and cyanobacteria because less energy is required to

incorporate and assimilate the former (Syrett, 1981; Gardner

et al., 2004; Flores and Herrero, 2005).

These issues were not addressed in recent studies sug-

gesting that eutrophication cannot be controlled by reducing

N inputs (e.g., Schindler et al., 2008; Wang and Wang, 2009).

The assumption that N2 fixing genera will replace non-N2

fixing genera like Microcystis when N is limiting and P is

sufficient could not be confirmed in Taihu. Furthermore, our

observation that Taihu does not fit the proposed “P only”

management paradigm of Schindler et al. (2008) is not unique.

Numerous other lakes, reservoirs, rivers and fjords worldwide

exhibit N and P co-limitation, either simultaneously or in

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3 1981

seasonally-shifting patterns (Dodds et al., 1989; Elser et al.,

2007; Forbes et al., 2008; Scott et al., 2008; North et al., 2007;

Lewis and Wurtsbaugh, 2008; Conley et al., 2009; Xu et al.,

2010; Abell et al., 2010).

4. Conclusions

Nutrient loading analyses, nutrient addition bioassays and

nutrient cycling studies provide the basis for recommending

that N control be included, along with the previously

prescribed P control (Chen et al., 2003a,b; Wang and Wang,

2009), as a nutrient management strategy for Taihu. Denitri-

fication rates, while high relative to other lakes, are lower than

estimates of N loading and therefore would not mitigate high

N loads. Also, late summer cyanobacterial blooms are main-

tained primarily by water column N regeneration. Recycling

produces NH4þ available to non-N-fixing cyanobacterial

blooms (Microcystis), regardless of the N form discharged into

the lake.

The fact thatMicrocystis spp. were not replaced by N2 fixing

cyanobacterial bloom species during N limited, but P sufficient

summer periods is evidence that predictions of succession

from non-N2 to N2 fixing taxa based on N:P stoichiometry

(Smith, 1990; Schindler et al., 2008) may not apply to hyper-

eutrophic lakes. Excess inputs of both N and P, combined with

internal cycling of these nutrients, may overwhelm the ability

of a single nutrient to control increasing eutrophication and

bloom intensification in Lake Taihu and other large lakes

experiencing such blooms (e.g., Lake Erie, Lake Okeechobee,

Lake Victoria).

P input reductions are an important component of eutro-

phication management in large lakes and reservoirs.

However, failure to control N inputs may result in continued

serious eutrophication problems caused by non-N2 fixing

cyanobacterial blooms.

Acknowledgements

We thank the TLLER, the Taihu Basin Authority, and the

Chinese Ministry of Water Resources for providing water

quality data, and XiaodongWang, Linlin Cai, Jingchen Xue, Lu

Zhang and Longyuan Yang for assistance with sampling and

chemical analyses and Guangbai Cui and Yong Pang for data

collection. This research was supported by the Chinese

Academy of Sciences (Contract: KXCX1-YW-14), the Ministry

of Science and Technology of China (Contract: 2009ZX07101-

013), the Chinese National Science Foundation (Contract:

40825004, 40730529, 51009049), Fundamental Research Funds

for the Central Universities (China), the US Environmental

Protection Agency (Project 83335101-0), and US National

Science Foundation (CBET Program) Project 0826819.

Appendix. Supplementary data

Supplementary data related to this article can be found online

at doi:10.1016/j.watres.2010.09.018

r e f e r e n c e s

Abell, J.M., Ozkundakci, D., Hamilton, D.P., 2010. Nitrogen andphosphorus limitation of phytoplankton growth in NewZealand lakes: implications for eutrophication control.Ecosystems. doi:10.1007/s10021-010-9367-9.

Ahn, C.-Y., Chung, A.-S., Oh, H.-M., 2002. Rainfall, phycocyanin,and N:P ratios related to cyanobacterial blooms in a Koreanlarge reservoir. Hydrobiologia 474, 117e124.

Andersen, J.M., 1975. Influence of pH on release of phosphorusfrom lake sediments. Archiv fur Hydrobiologie 76, 411e419.

APHA., 1995. Standard Methods for the Examination of Water andWastewater, nineteenth ed. American Public HealthAssociation, American Water Works Association, WaterEnvironment Federation.

Blackburn, T.H., 1979. Method for measuring rates of NH4þ

turnover in anoxic marine sediments, using a 15NeNH4þ

dilution technique. Applied & Environmental Microbiology 37,760e765.

Blomqvist, P., Pettersson, A., Hyenstrand, P., 1994.Ammoniumenitrogen: a key regulatory factor causingdominance of non-nitrogen-fixing cyanobacteria in aquaticsystems. Archives of Hydrobiology 132 (2), 141e164.

Boyer, E.W., Howarth, R.W., Galloway, J.N., Dentener, F.J.,Cleveland, C., Asner, G.P., Greene, P., Vorosmarty, C., 2004.Current nitrogen inputs to world regions. In: Mosier, A.R.,Syers, J.K., Freney, J.R. (Eds.), Agriculture and the NitrogenCycle. Island Press, Washington, DC, pp. 221e230. SCOPE #65.

Caperon, J., Schell, D., Hirota, J., Laws, E., 1979. Ammoniumexcretion rates in Kaneohe Bay, Hawaii, measured by a 15Nisotope dilution technique. Marine Biology 54, 33e40.

Carmichael, W.W., 2001. Health effects of toxin producingcyanobacteria: the cyanoHABs. Human and Ecological RiskAssessessment 7, 1393e1407.

Chen, Y.W., Qin, B.Q., Teubner, K., Dokulil, M.T., 2003a. Long-termdynamics of phytoplankton assemblages: Microcystis-domination in Lake Taihu, a large shallow lake in China.Journal of Plankton Research 25, 445e453.

Chen, Y.W., Fan, C.X., Teubner, K., Dokulil, M., 2003b. Changes ofnutrients and phytoplankton chlorophyll-a in a large shallowlake, Taihu, China: an 8-year investigation. Hydrobiologia 506,273e279.

Conley, D.J., Paerl, H.W., Howarth, R.W., Boesch, D.F.,Seitzinger, S.P., Havens, K.E., Lancelot, C., Likens, G.E., 2009.Controlling eutrophication: nitrogen and phosphorus. Science323, 1014e1015.

Dodds, W.K., Johnson, K.R., Priscu, J.C., 1989. Simultaneousnitrogen and phosphorus deficiency in naturalphytoplankton assemblages: theory, empirical evidence andimplications for lake management. Lake and ReservoirManagement 5, 21e26.

Ebina, J., Tsutsui, T., Shirai, T., 1983. Simultaneous determinationof total nitrogen and total phosphorus in water usingperoxodisulfate oxidation. Water Research 17, 1721e1726.

Elser, J.J., Bracken, M.E.S., Cleland, E.E., Gruner, D.S., Harpole, W.S.,Hillebrand, H., Bgai, J.T., Seabloom, E.W., Shurin, J.B., Smith, J.E.,2007. Global analysis of nitrogen and phosphorus limitation ofprimary producers in freshwater, marine and terrestrialecosystems. Ecololgy Letters 10, 1124e1134.

Ferber, L.R., Levine, S.N., Lini, A., Livingston, G.P., 2004. Docyanobacteria dominate in eutrophic lakes because they fixatmospheric nitrogen? Freshwater Biology 49, 690e708.

Flores, E., Herrero, A., 2005. Nitrogen assimilation and nitrogencontrol in cyanobacteria. Biochemical Society Transactions 33(1), 164e167.

Forbes, M.G., Doyle, R.D., Scott, J.D., Stanley, J.K., Huang, H.,Brooks, B.W., 2008. Physical factors control phytoplankton

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 31982

production and nitrogen fixation in eight Texas reservoirs.Ecosystems 11, 1181e1197.

Galloway, J.N., Cowling, E.B., 2002. Reactive nitrogen and theworld: 200 years of change. Ambio 16 (2), 64e71.

Gardner, W.S., Bootsma, H.A., Evans, C., St. John, P.A., 1995.Improved chromatographic analysis of 15N:14N ratios inammonium or nitrate for isotopic addition experiments.Marine Chemistry 48, 271e282.

Gardner, W.S., Lavrentyev, P.J., Cavaletto, J.F., McCarthy, M.J.,Eadie, B.J., Johengen, T.H., Cotner, J.B., 2004. The distributionand dynamics of nitrogen and microbial plankton in southernLake Michigan during spring transition 1999e2000. Journal ofGeophysical Research 109, 1e16.

Guo, L., 2007. Doing battle with the green monster of Lake Taihu.Science 317, 1166.

Havens, K.E., Fukushima, T., Xie, O.P., Iwakuma, T., James, R.T.,Takamura, N., Hanazato, T., Yamamoto, T., 2001. Nutrientdynamics and the eutrophication of shallow lakesKasumigaura (Japan), Donghu (PR China), and Okeechobee(USA). Environmental Pollution 111 (2), 262e272.

Howarth, R.W., Marino, R., Lane, J., Cole, J.J., 1988. Nitrogenfixation rates in freshwater, estuarine, and marineecosystems. Limnology and Oceanography 33, 669e687.

Hu, H., Li, Y., Wei, Y., Zhu, H., Shi, Z., 1980. Freshwater Algae inChina. Shanghai Science and Technology Press, Shanghai (inChinese).

Huisman, J.M., Matthijs, H.C.P., Visser, P.M., 2005. HarmfulCyanobacteria. Springer Aquatic Ecology Series 3. Springer,Dordrecht, The Netherlands, p. 243.

James, R.T., Havens, K., Zhu, G., Qin, B., 2009. Comparativeanalysis of nutrients, chlorophyll and transparency in twolarge shallow lakes (Lake Taihu, PR China and LakeOkeechobee, USA). Hydrobiologia 627, 211e231.

Jeppesen, E., Søndergaard, M., Meerhoff, M., Lauridsen, T.L.,Jensen, J.P., 2007. Shallow lake restoration by nutrient loadingreductiondsome recent findings and challenges ahead.Hydrobiologia 584, 239e252.

Kappers, F.I., 1980. The cyanobacterium Microcystis aeruginosa kg.and the nitrogen cycle of the hypertrophic Lake Brielle (TheNetherlands), pp. 37e43. In: Barica, J., Mur, L. (Eds.),Hypertrophic Ecosystems Dr. W. Junk (The Hague, TheNetherlands).

Kronvang, B., Jeppesen, E., Conley, D.J., Søndergaard, M.,Larsen, S.E., Ovesen, N.B., Carstensen, J., 2005. Nutrientpressures and ecological responses to nutrient loadingreductions in Danish streams, lakes and coastal waters.Journal of Hydrology 304, 274e288.

Lelieveld, J., Dentener, F., 2000. What controls troposphericozone? Journal of Geophysical Research 105, 3531e3551.

Lewis Jr., W.M., Wurtsbaugh, W.A., 2008. Control of lacustrinephytoplankton by nutrients: erosion of the phosphorusparadigm. Internationale Revue der gesamten Hydrobiologieund Hydrographie 93, 446e465.

Likens, G.E., 1972. In: Nutrients and Eutrophication Americ. Soc.Limnol. Oceanogr. Special Symp. 1.

McCarthy, M.J., Lavrentyev, P.L., Yang, L., Zhang, L., Chen, Y.,Qin, B., Gardner, W.S., 2007. Nitrogen dynamics relative tomicrobial food web structure in a subtropical, shallow, well-mixed, eutrophic lake (Taihu Lake, China). Hydrobiologia 581,195e207.

McCarthy, M.J., James, R.T., Chen, Y., East, T.L., Gardner, W.S.,2009. Nutrient ratios and phytoplankton community structurein the large, shallow, eutrophic, subtropical Lakes Okeechobee(Florida, USA) and Taihu (China). Limnology 10, 215e227.

North, R.L., Guildford, S.J., Smith, R.E.H., Havens, S.M., Twiss, M.R.,2007. Evidence for phosphorus, nitrogen, and iron colimitationof phytoplankton communities in Lake Erie. Limnology andOceanography 52, 315e328.

Ozkan, K., Jeppesen, E., Johansson, L.S., Beklioglu, M., 2009. Theresponse of periphyton and submerged macrophytes tonitrogen and phosphorus loading in shallow warm lakes:a mesocosm experiment. Freshwater Biology. doi:10.1111/j.1365e2427.2009.02297.x.

Paerl,H.W., 1990. Physiological ecologyandregulationofN2fixationin natural waters. Advances in Microbial Ecology 11, 305e344.

Paerl,H.W.,1988.Nuisancephytoplanktonbloomsincoastal,estuarine,and inlandwaters. Limnology and Oceanography 33, 823e847.

Paerl, H.W., 1997. Coastal eutrophication and harmful algalblooms: importance of atmospheric deposition andgroundwater as “new” nitrogen and other nutrient sources.Limnology and Oceanography 42, 1154e1165.

Paerl,H.W.,2008.Nutrientandotherenvironmentalcontrolsofharmfulcyanobacterial blooms along the freshwater-marine continuum.Advances in Experimental Medicine and Biology 619, 216e241.

Paerl, H.W., 2009. Controlling eutrophication along thefreshwateremarine continuum: dual nutrient (N and P)reductions are essential. Estuaries and Coasts 32, 593e601.

Paerl, H.W., Bowles, N.D., 1987. Dilution bioassays: theirapplication to assessments of nutrient limitation inhypereutrophic waters. Hydrobiologia 146, 265e273.

Paerl, H.W., Pinckney, J.L., 1996. Microbial consortia: their role inaquatic produc-tion and biogeochemical cycling. MicrobialEcology 31, 225e247.

Paerl, H.W., Fulton, R.S., Moisander, P.H., Dyble, J., 2001. Harmfulfreshwater algal blooms, with an emphasis on cyanobacteria.The Scientific World 1, 76e113.

Papista, E., Acs, E., Boeddi, B., 2002. Chlorophyll-a determinationwith ethanol e a critical test. Hydrobiologia 485, 191e198.

Piehler, M.F., Dyble, J., Moisander, P.H., Chapman, A.D.,Hendrickson, J., Paerl, H.W., 2009. Interactions betweennitrogen dynamics and the phytoplankton community in LakeGeorge, Florida, USA. Lake andReservoirManagement 25, 1e14.

Pu, P., Yan, J., 1998. Taihu Lake e a large shallow lake in the EastChina plain. Journal of Lake Sciences (China) 10 (suppl), 1e12.

Qin, B.Q.,Xu, P.Z.,Wu,Q.L., Luo, L.C., Zhang,Y.L., 2007. Environmentalissues of Lake Taihu, China. Hydrobiologia 581, 3e14.

Qin, B., Zhu, G., Gao, G., Zhang, Y., Li, W., Paerl, H.W.,Carmichael, W.W., 2010. A drinking water crisis in Lake Taihu,China: linkage to climatic variability and lake management.Environmental Management 45, 105e112.

Rabalais, N.N., 2002. Nitrogen in aquatic ecosystems. Ambio 16(2), 102e112.

Reynolds, C.S., 1987. Cyanobacterial water blooms. Advances inBotanical Research 13, 67e143.

Schindler, D.W., 1977. The evolution of phosphorus limitation inlakes. Science 195, 260e262.

Schindler, D.W., Hecky, R.E., Findlay, D.L., Stainton, M.P.,Parker, B.R., Paterson, M., Beaty, K.G., Lyng, M., Kasian, S.E.M.,2008. Eutrophication of lakes cannot be controlled by reducingnitrogen input: results of a 37 year whole ecosystemexperiment. Proceedings of the National Academy of ScienceUSA 105, 11254e11258.

Scott, J.T., Doyle, R.D., Prochnow, S.J., White, J.D., 2008. Arewatershed and lacustrine controls on planktonic N2 fixationhierarchically structured? Ecological Applications 18,805e819.

Scott, J.T., McCarthy, M.J., 2010. Nitrogen fixation may not balancethe nitrogen pool in lakes over timescales relevant toeutrophication management. Limnology and Oceanography55, 1265e1270.

Smith, V.H., 1983. Low nitrogen to phosphorus ratios favordominance by blueegreen algae in lake phytoplankton.Science 221, 669e671.

Smith, V.H., 1990. Nitrogen, phosphorus, and nitrogen fixation inlacustrine and estuarine ecosystems. Limnology andOceanography 35, 1852e1859.

wat e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 9 7 3e1 9 8 3 1983

Syrett, P., 1981. Nitrogen metabolism in microalgae. Physiologicalbases of phytoplankton ecology. Canadian Journal of Fisheriesand Aquatic Science 210, 182e210.

Vallino, J.J., Hopkinson, C.S., Hobbie, J.E., 1996. Modeling bacterialutilization of dissolved organic matter: optimization replacesMonod growth kinetics. Limnology and Oceanography 41,1591e1609.

Vitousek, P.M., Mooney, H.A., Lubchenko, J., Mellilo, J.M., 1997.Human omination of earth’s ecosystem. Science 277,494e499.

Wang, H.J., Wang, H.Z., 2009. Mitigation of lake eutrophication:loosen nitrogen control and focus on phosphorus abatement.Progress in Natural Science 19, 1445e1451.

Xu, H., Paerl, H.W., Qin, B.Q., Zhu, G.W., Gao, G., 2010. Nitrogen andphosphorus inputs control phytoplankton growth in eutrophicLake Taihu, China. Limnology and Oceanography 55, 420e432.

Zhai, S., Yang, L., Hu, W., 2009. Observations of atmosphericnitrogen and phosphorus deposition during the period of algalbloom formation in Northern Lake Taihu, China.Environmental Management 44, 542e551.