Effects of large-scale changes in emissions on nutrient concentrations in Estonian rivers in the Lake Peipsi drainage basin

Effects of large-scale changes in emissions on nutrient

concentrations in Estonian rivers in the Lake Peipsi drainage basin

Arvo Iitala,*, Per Stalnackeb, Johannes Deelstrac, Enn Loigua, Margus Pihlakd

aInstitute of Environmental Engineering, Tallinn Technical University, Ehitajate tee 5, 19086 Tallinn, EstoniabNIVA—Norwegian Institute for Water Research, P.O. Box 173, Kjelsas, N-0411 Oslo, Norway

cJordforsk-Norwegian Centre for Soil and Environmental Research, F. A. Dahls vei 20, N-1432 As, NorwaydInstitute of Mathematical Statistics, Tartu University, J. Liivi 2-513, 50409 Tartu, Estonia

Received 30 November 2003; revised 1 May 2004; accepted 1 July 2004

Abstract

The fall of the Iron Curtain resulted in dramatic changes in Eastern Europe, including substantial reductions in the use of

fertilisers and livestock production, as well as a marked decrease in water consumption by both the general population and

industries. This situation has created a unique opportunity to study the way that rivers have responded to these changes. Here, the

impact of these reductions on concentrations of nutrients (N and P) at 22 sampling sites on Estonian rivers are examined. There

were statistically significant downward trends (one-sided test at the 5% level) in total nitrogen (TN) concentrations at 20 of the 22

sites. These decreases in TN relate to: (i) substantial reductions in the use of organic and inorganic fertilisers, (ii) reduction of

cultivated and ploughed areas and increased proportions of grassland and abandoned land and (iii) improvements in farm

management practices. For total phosphorus (TP), significant downward trends were detected at only two sites, and there were

also two upward trends. The TP trends can be mainly explained by changes in phosphorus discharges from municipal sewage

treatment plants. Fifteen downward trends and one statistically significant upward trend were found for the TN:TP ratio. The

general decline in this ratio has likely been conducive to blue-green algae blooms in the recipient, Lake Peipsi.

q 2004 Elsevier B.V. All rights reserved.

Keywords: Land use; Nitrogen; TN:TP ratio; Phosphorus; Trend analysis; Water quality; Estonia

1. Introduction

Many field studies have shown that losses of

nitrogen in surface runoff are correlated with the rates

and methods of fertiliser application, the proportion of

0022-1694/$ - see front matter q 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.jhydrol.2004.07.034

* Corresponding author. Tel.: C372 6018059; fax: C372

6014406.

E-mail address: [email protected] (A. Iital).

land that is cultivated, and the prevailing farm

management practices (Kauppi, 1979; Loigu and

Velner, 1985; Rekolainen, 1989; Keeney and DeLuca,

1993; Zablocki and Pienkovski, 1999; Mander et al.,

1998, 2000; Tumas, 2000; Kutra et al., 2002). The

studies mainly dealt with the possible impact of

increased emissions of nutrients on the water quality.

During the last 10–15 years considerable changes in

agricultural practice and waste water treatment took

Journal of Hydrology 304 (2005) 261–273

www.elsevier.com/locate/jhydrol

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273262

place in Eastern Europe, but only a few river-basin

scale studies examined the impact on rivers exerted by

the less extensive land-use and decreases in fertiliser

application combined with the improved performance

of wastewater treatment plants (WWTP). The results

are widely varying and both lack of or weak

responses in rivers (Berankova and Ungerman, 1996;

Prochazkova et al., 1996; Tonderski, 1997; Stalnacke

et al., 2003; Povilaitis et al., 2003) as well as strong

downward trends for nutrient concentrations have been

reported (Pekarova and Pekar, 1996; Olah and Olah,

1996; Hussian et al., in press; Povilaitis et al., 2003).

An unprecedented decrease in the use of fertilisers

and livestock production in Estonia that has been

greater than in most other Eastern European countries

(Fertilizer Yearbook, 2002), as well as the implemen-

tation of more effective waste water treatment

technologies in municipalities and industries has

created a unique opportunity to study the way that

rivers have responded to these changes.

Inorganic fertilisers were widely used in Estonian

agriculture during the Soviet period (i.e. before 1991),

and application reached a peak of 270,000 tons

(300 kg haK1) annually in 1987–1988 (Agriculture in

Estonia, 1997). However, comprehensive economic,

technical and social changes took place in Estonia after

the country regained its independence. For example,

the use of fertilisers has decreased considerably over

the last 15 years, and the levels observed in 2001

constituted only about 11% (29,700 tons) of the peak

in 1987–1988 that correspond to applications of less

than 100 kg haK1 for mineral fertilisers (63 kg N haK1

and 13 kg P2O5 haK1) and less than 30 tons haK1

for organic fertilisers (Agriculture, 2001, 2002).

Furthermore, the number of livestock units decreased

from 800,000 (0.82 LU haK1 of arable land) in 1988 to

less than 390,000 (0.34 LU haK1 of arable land) in

1994, and the level today is approximately the same as

in 1994 (Agriculture, 2001, 2002). Slaughtering of

livestock reduced correspondingly the amount of

manure.

Point source emissions of N and P to surface waters

in Estonia have also decreased as a result of the

economic recession in the early 1990s. This was due

to a combination with modernisation of industrial

production and the construction of new and improve-

ment of existing wastewater treatment plants

(WWTPs) in major towns (Vassiljev et al., 2001).

Studies carried out in some smaller agricultural

watersheds in Estonia have shown relatively low

levels of N and P, as compared to the concentrations

found in, for example, the Nordic countries (Loigu

and Iital, 2000; Iital and Loigu, 2001; Vagstad et al.,

2001, Iital et al., 2002). The low levels of N and P

detected in Estonia might be explained by the

following: (i) decreased rates of fertiliser application;

(ii) lower livestock density; (iii) differences in land

use (more natural and cultural grasslands); (iv)

hydrological conditions that entail longer water

residence time and higher buffering capacity, and

substantial retention of these substances within

catchments. The majority of the field drainage ditches

used today were established in the 1970s and 1980s,

and maintenance of the main ditches in these land

improvement systems has been insufficient in recent

years. Due to this situation, many watercourses are

now overgrown with bushes and macrophytes, which

has enhanced the retention potential, denitrification

and biological uptake.

This study, carried out in the Lake Peipsi drainage

basin that makes up 36% of the Estonian territory,

investigates the effects of large-scale changes in

emissions on nutrient concentrations and how long it

will take to detect the response of a river system to

changes in land use or agricultural practices. The time

series of nitrogen and phosphorus concentrations

measured in rivers are statistically evaluated and

special attention was paid to natural variation in

runoff–concentration relationship. That kind of infor-

mation can help environmental authorities and decision

and policy makers establish realistic goals. Such

knowledge can also be highly useful in the implemen-

tation of river basin management plans within the EU

Water Framework Directive and in selection of

adequate measures to achieve good ecological and

chemical status of all water bodies by 2015.

2. Study area, data base and methods

2.1. Study area

The Lake Peipsi catchment, including the surface of

the lake itself, has an area of 47,800 km2, of which

16,323 km2 is in Estonia and the rest in Russia and

Latvia. The northern part of this drainage basin has a

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 263

sedimentary cover consisting of Ordovician and

Silurian limestones; the southern part is characterised

by sandy-silty and clayey Devonian deposits, which

are overlaid by quaternary deposits that are usually less

than 5 m thick, and are thickest (often O100 m) in the

uplands. The topography of the catchment is relatively

flat, with maximum elevations of about 30–100 m

above sea level. The glaciolacustrine or till-covered

plain of the Estonian part of the Peipsi catchment is

higher in the south-west and north-west due to the

presence of the Pandivere (166 m), Haanja (318 m),

Otepaa (217 m), and Sakala (146 m) uplands. Most

rivers that flow into Lake Peipsi have their sources in

these elevated areas. Lake Peipsi drainage area is a sub-

catchment of the basin of the Gulf of Finland and the

Baltic Sea, and Lake Peipsi is connected with the basin

of the Gulf of Finland and the Baltic Sea via the Narva

River. The mean air temperature in the Peipsi

catchment is 14–15 8C in June and K4 to 4.5 8C in

December, and the mean annual precipitation is 600–

650 mm (Jaagus and Tarand, 1988). Forests predomi-

nate land cover in the north of the catchment (up to 60–

70%), but coverage decreases southwards (about 30–

40%). The proportion of mires is higher in the drainage

areas of rivers in the northern part of the region.

Agriculture in the Lake Peipsi catchment includes

Fig. 1. Map showing th

animal husbandry and the production of arable crops,

mainly cereals.

2.2. Sampling strategy and analytical methods

Statistical analysis was undertaken to discern

trends in time series of total nitrogen (TN) and

total phosphorus (TP) concentrations and TN:TP

ratios for 22 sampling sites on 17 rivers and streams

in Estonia. A monthly time resolution was used for

data from 18 of the sampling sites, and a bimonthly

resolution for the remaining four sites (i.e. Poltsa-

maa-Rutikvere, Vohandu-Himmiste, Pedja-Torve,

and Ohne-Rooba; Fig. 1). To analyse total nitrogen,

samples were digested with peroxodisulphate, after

which the concentrations of nitrate were determined.

Total phosphorus was analysed by using the

peroxodisulphate digestion procedure to convert the

various forms of phosphorus into dissolved orthopho-

sphate. The time series covered at least the period

1986–2001 for 14 of the sites, whereas only 1991 (or

1992) to 2001 was investigated for the other eight

sites (Table 2). Both small and large sub-catchments

(ranging in size from 108 to 47,815 km2) were

studied in all parts of the Estonian portion of the

Lake Peipsi basin (Table 1). The drainage areas of

the Narva (site No. 19) and Piusa (site No. 9) Rivers

e sampling sites.

Table 1

Main characteristics of the rivers monitored in the Lake Peipsi drainage basin

Site no. River Name of sampling

site

Catchment

area (km2)

Population

density

(inhab/km2)

Land type

Forest (%) Wetlands,

other natural

(%)

Agricultural

(%)

Arable

(%)

1 Alajogi Alajoe HS 140 2.7 67.2 17.5 14.9 2.5

2 Ahja Kiidjarve 336 10.9 44.9 3.2 51.3 10.4

3 Poltsamaa Rutikvere 861 15.8 45.5 10.0 42.4 20.7

4 Rannapungerja Roostoja 214 5.0 59.5 17.5 21.1 7.8

5 Emajogi Rannu-Joesuu HS 3,374 16.2 42.4 6.0 41.9 15.6

6 Vohandu Rapina 1,144 29.8 43.4 6.4 47.1 9.5

7 Vohandu Himmiste HS 848 32.9 45.0 4.2 47.4 10.5

8 Tanassilma Oiu 454 31.9 46.2 10.3 42.7 20.8

9 Piusa Korela 523 9.5 48.9 4.5 45.7 7.2

10 Avijogi Mulgi HS 366 8.2 66.2 7.0 26.1 11.4

11 Tagajogi Tudulinna 252 2.8 72.7 20.5 6.4 1.7

12 Kaapa Kose 282 5.1 60.5 10.5 27.9 15.9

13 Tarvastu Podraoja 108 16.4 41.8 0.9 56.3 30.9

14 Pedja Jogeva SAJ 665 7.5 57.5 7.7 33.9 19.9

15 Pedja Torve HS 776 15.7 55.8 7.9 34.8 21.0

16 Vaike-Emajogi Pikasilla 1,270 24.1 48.7 4.0 44.5 12.5

17 Emajogi Tartu HS 7,828 9.4 45.8 10.4 41.6 21.3

18 Emajogi Kavastu 8,539 21.9 44.0 9.5 43.6 20.9

19 Narva Vasknarva HS 47,815 6.9 47.0 9.6 39.3 15.6

20 Ohne Suislepa 557 14.3 49.3 9.5 38,9 15.4

21 Ohne Roobe 266 5.4 56.1 14.2 27.6 13.8

22 Porijogi Reola HS 241 12.9 41.8 2.5 55.0 12.5

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273264

also include territory within the Russian Federation

(69 and 34%, respectively).

2.3. Statistical methods

The statistical properties of water quality data

(nutrient concentrations) are usually not normally

distributed, and they often exhibit a seasonal pattern

because they are influenced by water discharge

(Gilliom and Helsel, 1986). Here, a recently modified

version of the seasonal Mann–Kendall test (Libseller

and Grimvall, 2002), referred to as the partial Mann–

Kendall (PMK) test, which has been adapted to

account for the influence of confounding (i.e.

meteorological or hydrological) variables was used

with water discharge as such a variable. The

univariate Mann–Kendall statistic for a time series

{Zk, kZ1,2,.,n} of data is defined as

T ZXj!i

sgnðZi KZjÞ

where

sgnðxÞ Z

1; if xO0

0; if x Z 0

K1; if xO0

8><>:

If there are no ties between the observations, and

there is no trend in the time series, the test statistic is

asymptotically normally distributed with

EðTÞ Z 0 and VarðTÞ Z nðn K1Þð2n C5Þ=18

If the response variable is measured during several

(u) seasons, the seasonal Mann–Kendall test is

computed by first separating the data into u subseries,

each of which represents a season. In this way

Tj ZXk!l

signðZlj KZkjÞ j Z 1;.;u

is the Mann–Kendall statistic for season j, which is

summed over all seasons to obtain the seasonal

statistics

Fig. 2. Long-term mean annual discharge measured from 1978 to

2000 in the Alajogi River at the Alajoe sampling site.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 265

S ZXu

jZ1

Tj

The Mann–Kendall statistic (MK-Stat) is the

function of S, and this statistic has a standard normal

distribution. The conditional mean of the test statistic is

EðT 0rjT

0j Þ Z VarðTjrÞT =VarðTjÞ

and variance

VarðT 0rjT

0j Þ Z VarðT 0

rÞKVarðTjrÞ2=VarðTÞ

where T 0r is a statistic for the response variable, T 0

j is a

statistic for the explanatory variable, VarðT 0rÞ and

VarðT 0j Þ, respectively, denote the variance of T 0

r and T 0j ,

and VarðTjrÞ is the covariance between the test statistic

of the response variable and the covariate. If there is

only one response (nutrient concentrations) and one

explanatory variable (water discharge), the distri-

bution of the test statistic will be asymptotically

normal.

Fig. 3. Relationship between TN concentrations (0.9 percentile) in

rivers and the share of arable land in the catchment in 1987–2001.

3. Results

The trends in average annual water discharge at the

studied sampling sites during the monitored period

show rather remarkable variability. According to the

data from the Estonian Meteorological and Hydro-

logical Institute, the average annual discharge

increased in the Avijogi and Tagajogi Rivers, but

decreased slightly in the Alajogi, Kaapa, Tarvastu,

Poltsamaa, Vohandu, Emajogi, and Pedja Rivers

(Fig. 2). For some rivers (the Ohne, Vaike-Emajogi,

Narva, and Ahja), no trends in long-term mean annual

discharge were detected.

The proportion of agricultural land (arable land and

permanent natural grasslands) in the sub-catchments

is still relatively high today, in most cases at least

40%. However, the share of arable land (temporary

crops, temporary meadows, land temporarily in

fallow and abandoned land) is generally rather small

(Table 1), and the mean TN concentrations in the

rivers were strongly correlated to the fraction of arable

land in the catchment (Fig. 3). The population

densities were highest (more than 30 inhabitants

kmK2) in the drainage areas of the Vohandu and

Tanassilma Rivers (Table 1) and were fairly low (less

than 3 inhabitants per km2) in the northern part of the

Lake Peipsi drainage basin, more precisely in the

drainage areas of the Alajogi and Tagajogi Rivers,

which are also characterised by large forested and

wetland areas (O85%).

TN concentrations in the rivers showed downward

trends at almost all sites (Table 2). Very rapid

decrease in TN levels occurred as early as the

beginning of the 1990s at some of the locations (e.g.

Vohandu-Rapina, Avijogi-Mulgi, Ahja-Kiidjarve, and

Poltsamaa-Rutikvere; Figs. 4 and 5). Significant

downward trends in TN were also observed in some

of the sub-catchments dominated by forests and

wetlands. At almost all sites there was the occurrence

of pronounced seasonal fluctuation in TN, with the

lowest levels in the summer period and the highest

concentrations from late autumn to spring. The PMK

test revealed statistically significant (p!0.05; one-

sided test) downward trends in TN at 20 of the 22

sampling stations (Table 2), and the level of

significance was high (p!2%) at 18 of these 20

sites. No significant upward trends were observed at

any of the sampling locations.

Table 2

MK statistics and long-term trends for nutrient concentrations in the rivers in the Lake Peipsi catchment

Station Sampling site Total N Total P

Years

monitored

MK-Stat p-valuea Years

monitored

MK-Stat p-valuea

1 Alajogi-Alajoe 1984–2001 K1.85 0.032 1984–2001 2.20 0.014

2 Ahja-Kiidjarve 1985–2001 K3.85 !0.001 1981–2001 K1.20 0.114

3 Poltsamaa-Rutikvere 1986–2001 K3.71 !0.001 1986–2001 K1.34 0.091

4 Rannapungerja-Roostoja 1992–2001 0.06 0.475 1992–2001 K1.41 0.079

5 Emajogi-Joesuu 1987–2001 K3.43 !0.001 1987–2001 0.87 0.193

6 Vohandu-Rapina 1986–2001 K3.92 !0.001 1986–2001 K0.74 0.231

7 Vohandu-Himmiste 1991–2001 K2.76 0.003 1986–2001 K1.30 0.098

8 Tanassilma-Oiu 1986–2001 K3.59 !0.001 1986–2001 K3.75 !0.001

9 Piusa-Korela 1992–2001 K2.35 0.009 1992–2001 K0.29 0.386

10 Avijogi-Mulgi 1992–2001 K2.25 0.012 1992–2001 K1.31 0.094

11 Tagajogi-Tudulinna 1992–2001 K2.26 0.012 1992–2001 0.27 0.393

12 Kaapa-Kose 1986–2001 K2.30 0.011 1986–2001 K0.07 0.472

13 Tarvastu-Podraoja 1986–2001 K3.35 !0.001 1986–2001 K2.61 0.005

14 Pedja-Jogeva 1986–2001 K0.01 0.495 1987–2001 K0.83 0.202

15 Pedja-Torve 1986–2001 K2.51 0.006 1987–2001 K0.51 0.304

16 Vaike-Emajogi-Pikasilla 1986–2001 K3.79 !0.001 1986–2001 K0.27 0.393

17 Emajogi-Tartu 1984–2001 K2.06 0.020 1982–2001 1.02 0.154

18 Emajogi-Kavastu 1986–2001 K1.75 0.040 1986–2001 K0.86 0.196

19 Narva-Vasknarva 1992–2001 K1.87 0.031 1992–2001 1.05 0.146

20 Ohne-Suislepa 1986–2001 K3.58 !0.001 1986–2001 2.31 0.010

21 Ohne-Roobe 1991–2001 K2.36 0.009 1986–2001 1.29 0.099

22 Porijogi-Reola 1992–2001 K2.99 0.001 1992–2001 0.04 0.485

a The p-value is significant if !0.05. Bold type indicates results that are statistically significant at the 0.05 level (one-sided test).

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273266

For TP, the results of the PMK test were

statistically significant for data from four sites

(Table 2): two downward trends (Tanassilma and

Tarvastu) and two upward trends (Alajogi, Ohne-

Suislepa).

Analysing the TN:TP mass ratios, 15 statistically

significant downward trends and one upward trend

was detected. Six of the 18 rivers showed no trends

in TN:TP mass ratios, and five of those six (i.e. the

Ahja, Poltsamaa, Tarvastu, Pedja-Torve, and Narva)

exhibited significant downward trend in TN

concentrations.

4. Discussion

Several studies conducted in the last decade have

examined the impact on rivers exerted by the rapid

changes in land-use and decreases in fertiliser

application in Eastern Europe. Notably, a review

prepared by Stalnacke et al. (2004) showed that

widely varying results have been reported for the

different areas that have been studied. For example,

lack of (or only weak) responses in rivers (i.e. no

downward trends) have been reported for N and P

concentrations by Berankova and Ungerman (1996),

Prochazkova et al. (1996) and Tonderski (1997), and

for nitrogen levels by Stalnacke et al. (2003) and

Povilaitis et al. (2003). Furthermore, downward trends

in Eastern European rivers have been observed for

nutrients by Pekarova and Pekar (1996), Olah and

Olah (1996) and Hussian et al. (2004), and for

phosphorus in particular by Stalnacke et al. (2003)

and Povilaitis et al. (2003).

In Estonia, studies of this kind have previously

been performed chiefly in single, smaller catchments.

Loigu and Vassiljev (1997) noted a clear downward

trend in nitrate concentrations in the rivers of the

small and intensively cultivated Kurna catchment

(23.2 km2, 68% arable land) during the period 1987–

1996. This trend was particularly conspicuous for the

peak concentrations, and the cited authors stated that

Fig. 4. Time series of total nitrogen concentrations in rivers in the Lake Peipsi basin.

Fig. 5. Variation in concentrations of total nitrogen (TN) at the

Vohandu-Rapina sampling station in 1986–2001.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 267

it could be explained by the decreased application of

fertilisers and agricultural production in the catch-

ment. Similarly, Mander et al. (2000) found down-

ward nitrogen and phosphorus trends in another small

(258 km2) agricultural catchment in Estonia.

The results of these earlier investigations con-

ducted in Estonia are confirmed by the findings of our

comprehensive study, which represent both small and

large sub-catchments of the entire Estonian part of the

Lake Peipsi drainage area. Our most notable obser-

vations are as follows: (i) the decrease in TN

concentrations was rapid, occurring as early as the

beginning of the 1990s at many of the sites (Fig. 1);

Fig. 6. Consumption of mineral fertilisers in Estonia in 1990–2001.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273268

(ii) there was also a decrease in TN concentrations at

sites with catchments dominated by forests; (iii) there

was very little evidence of downward trends in TP

concentrations; (iv) there was a decline in the

amplitude and total variability of the nitrogen series

between the 1980s and the early/mid 1990s, after

which, the pattern of amplitude and variability of the

TN concentrations was rather stable. These interesting

findings require more detailed discussion.

It is widely accepted that nutrient concentrations in

streams and rivers may respond differently to changes

in physical–geographical conditions, agricultural

production, rates of fertiliser application, and intensity

of land use. Plot experiments carried out in small

watersheds in Estonia during the late 1990s have

shown that substantial amounts of nitrogen can be lost

via the root zone after the harvesting of crops,

resulting in nitrate concentrations as high as 60 mg

N lK1 in soil–water (Loigu and Iital, 2000; Tamm,

2001). At the same time, the nitrate concentrations in

groundwater and small agricultural streams are low,

only about 2–4 mg N lK1. Stalnacke et al. (1999)

pointed out that denitrification, presumably in smaller

streams and channels, plays an important role in

nitrogen reduction in the Baltic countries. The pH in

stream water is usually high (R8) during summer,

which promotes the volatilisation of ammonia. In

addition, the water residence times in Estonian river

catchments are much longer than in many other areas

with similar geographical conditions (Deelstra, et al.,

1998). All of the cited studies clearly indicate that a

plot-field catchment system has a considerable

potential for nitrogen retention. Assuming that the

retention was also high in the 1980s, when the levels

of fertiliser application and agricultural intensity were

greatest, the rapid decline in TN concentrations that

occurred in Estonia during the early 1990s is some-

what surprising. In Latvia, where there was a similar

drop in fertiliser use and the hydro-meteorological

conditions are similar to those in Estonia, only weak

downward trends in dissolved inorganic nitrogen were

reported for the same time period (Stalnacke et al.,

2003). This can, in part, be explained by the larger

size of the river catchments studied in Latvia and by

differences in hydro-geological conditions (i.e. poss-

ible nitrogen retention in groundwater systems). In

our study, the downward trends in TN in forested

catchments (e.g. Tagajogi, Alajogi, and Avijogi) were

probably caused primarily by the same factors that

induced the downward trends in other catchments:

namely, insufficient maintenance of main ditches

during the past decade, resulting in overgrowth of

bushes and macrophytes, which in turn enhances the

retention potential (i.e. denitrification and biological

uptake).

The improvement of agricultural management

practices in Estonia during the 1990s has led to

more sustainable use of inorganic and organic

fertilisers. This change has resulted in a remarkable

decrease in the maximum concentrations of nitrogen

in rivers (Loigu and Vassiljev, 1997), because, for

example, during the Soviet period mineral fertilisers

and manure were applied also to frozen or snow

covered soil. According to FAO Statistics (Fertilizer

Yearbook, 2002), the decline in the use of fertilisers,

especially nitrogenous fertilisers, has been much

greater in Estonia (Fig. 6) than in most other Eastern

European countries. Considering the area of arable

land (crop fields and cultural grasslands) in Estonia,

about 34% was unused in 2001 (Agriculture, 2001,

2002), and the share of wintergreen area was quite

remarkable, being as high as 80% in some small

agricultural catchments. Moreover, it is possible that

the particular hydro-geological conditions influence

the rate at which nutrient concentrations in streams

and rivers respond to changes in land use and nutrient

emissions. The Estonian bedrock consists largely of

Silurian and Ordovician limestone. Hence, there can

be rapid exchange of water between upper and lower

water tables through existing cracks. Many rivers,

especially in the Pandivere Upland, are fed by

groundwater that is rich in nitrate-nitrogen. After the

fall of the Soviet Union, the nitrate content in

groundwater aquifers decreased substantially

(Tamm, 2003). Several field studies have shown that

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 269

at locations where oxygen present at very low

concentrations and where for example organic carbon

is available significant denitrification has been found.

Gambrell et al. (1975) concluded that losses of nitrate-

nitrogen in undisturbed, poorly drained soils with

relatively high water tables are lower compared to

naturally well-drained soils, mainly as a result of

denitrification. McMahon and Bohlke (1996) studied

denitrification capacity in Nebraska’s South Platte

River aquifer which is affected by irrigation, and

found that denitrification accounted for 15–30% of the

decrease in NO3 concentrations. Gilliam and Skaggs

(1986) and Evans et al. (1992) have shown that water

table management can reduce NO3–N in drainage

water by over 60%. This decrease was at least partly

caused by an increase in denitrification. Due to

insufficiently maintained amelioration systems, the

groundwater level rose in Estonia in 1990s and the

soil is maintained at saturated or almost saturated

conditions for longer time periods leading to anaero-

bic environments in soils, possible increase in

denitrification rates and to lower TN concentrations

in open streams. The significant downward trend in

TN concentrations at site No. 19 (Narva-Vasknarva),

which actually describes the quality of the water in

Lake Peipsi, was probably caused by the marked

decrease in nitrogen loads (Blinova, 2001) transported

to the lake during the last decade. Seasonal trends

were also investigated here, but the overall picture

provided by our results was not clear on many

occasions, probably due to the low TN concentrations.

However, at some of the sites with a high proportion

of agricultural or arable land (i.e. Tanassilma,

Vohandu-Rapina, Piusa, and Porijogi), significant

downward trends were more pronounced during

autumn (September to November) and winter

(December to February). Seasonal trends were not

as significant at the other sites with a high share of

arable land (e.g. Poltsamaa, Tarvastu, and Pedja-

Torve). Significant trends in TN concentrations were

not detected in the upstream part of the Pedja River

(Pedja-Jogeva sampling site), which is an area that is

generally not directly affected by human activities.

The present statistical trend analysis clearly indi-

cated that the concentrations of TP in most rivers were

fairly stable throughout the periods studied. Moreover,

other investigators (Loigu, 1993; Haraldsen et al.,

2001) have found that the amount of P in the topsoil in

Estonia is still at the same level as in the late 1980s,

even though there had been a substantial decrease in

the rate of application of P fertilisers. The soil has a

large capacity to absorb phosphorus, but release of that

element can occur after a time lag of several years

(Vagstad et al., 2001). A clear downward trend in TP

levels was observed in only two rivers, both draining

into Lake Vortsjarv. This can probably be explained by

more efficient treatment of municipal wastewater, a

decrease in the number of inhabitants in rural areas,

and, in particular, less intensive agriculture. In

addition, the two rivers that showed downward trends

in phosphorus have relatively small catchments, which

is probably the reason that responses to the decrease in

loads of this element were more readily detected.

Despite the improved performance of WWTPs the

overall phosphorus load to the rivers is still fairly high.

The reason for this is that the number of households and

industries connected to the sewerage systems in larger

municipalities has grown, which has increased the

amount of nutrient-laden wastewater delivered to the

treatment plants. Furthermore, rises in the price of

drinking water and wastewater treatment have led to a

substantial drop in the overall consumption of potable

water by the general population, which has resulted in

increased concentrations of nutrients in both the

wastewaters delivered to WWTPs and the effluents

released from those facilities. A typical example of this

can be seen in the city of Tartu, which represents the

largest single source of pollution in the study area. The

efficiency of phosphorus removal is about 80% at the

sewage treatment plant in Tartu. However, due to high

levels of this element in inlet waters (on average

12.5 mg P lK1 in 2002), the concentrations in the

effluent have exceeded the maximum permissible level

of 1 mg P lK1 (Fig. 7). In some sub-catchments with

high relative point source emissions, evidence of

decreased phosphorus concentrations was found, and

this was particularly noticeable during high-flow

periods due to a dilution effect. For instance, in the

Tanassilma River (site No. 8), the observed downward

trend in TP was most likely related to the decline in

the amount of wastewater delivered to the WWTPs

(Fig. 8). Considering the 0.9 percentile of TP

concentrations, there was a decrease from

0.18 mg lK1 in the early 1990s to 0.11 mg lK1 in

2001. Only the Tanassilma and Tarvastu Rivers

exhibited downward trends in concentrations of both

Fig. 7. Concentration of total phosphorus (TP) in the effluent water discharged from the wastewater treatment plant in Tartu in 2001 and 2002.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273270

TN and TP. However, the Tanassilma was also the only

river in which an upward trend in the TN:TP mass ratio

was detected. These findings indicate that the TP

concentrations have decreased more than the corre-

sponding TN concentrations. It is assumed that

phosphorus compounds limit primary production and

therefore also control the trophic level in surface

waters (Jarvekulg, 1993). Consequently, a low N:P

ratio may increase the probability of nitrogen limi-

tation in surface waters. In our study, a decline in the

N:P ratios was noted at 15 of the 22 sites we

investigated (Table 3, Fig. 9). Moreover, other

researchers (Noges et al., 2002) have observed down-

ward trends in N:P ratios and more intensive

cyanobacteria blooms in Lake Peipsi in the 1990s,

and the latter phenomenon was particularly pro-

nounced in 2001 and 2002. Given that cyanobacteria

can directly utilise molecular nitrogen, reducing

phosphorus emissions will no doubt improve

Fig. 8. Time series of phosphorus concentration

the quality of surface water in the Lake Peipsi

catchment. On the other hand, the economic recovery

that is expected in Estonia in coming decades will

probably also lead to more intensive agriculture (e.g. a

larger proportion of cultivated land and a higher rate of

fertiliser application) and consequently increase the

losses of nutrients to waters, especially with respect to

nitrogen (Mourad et al., 2003).

5. Conclusions

†

s in

Twenty statistically significant downward trends in

TN (one-sided test at the 5% level) were found

from a total of 22 sites on rivers in Estonia, and it is

obvious that the rivers have responded to the

following: (i) a dramatic decrease in the use of

organic and inorganic fertilisers and livestock

numbers; (ii) increased proportions of grassland

the Tanassilma River in 1992–2002.

Fig

200

Table 3

Results of univariate Mann–Kendall testing of N:P mass ratios for rivers in the Lake Peipsi catchment

River/sampling site Univariate MK test, N/P River/sampling site Univariate MK test, N/P

Alajogi/Alajoe K2.16 Kaapa/Kose K1.78

Ahja/Kiidjarve K0.10 Tarvastu/Podraoja K0.93

Poltsamaa/Rutikvere 0.59 Pedja/Jogeva K2.70

Rannapungerja/Roostoja K0.63 Pedja/Torve K1.45

Emajogi/Joesuu K2.58 Vaike-Emajogi/ Pikasilla K3.12

Vohandu/Rapina K3.46 Emajogi/Tartu K2.44

Vohandu/Himmiste K2.03 Emajogi/Kavastu K1.85

Tanassilma/Oiu 2.09 Narva/Vasknarva K1.66

Piusa/Korela K2.71 Ohne/Suislepa K3.48

AvijogiMulgi K1.79 Ohne/Roobe K2.28

Tagajogi/Tudulinna K2.09 Porijogi/Reola K2.55

Results given in bold type are statistically significant at the 0.05 level (one-sided test).

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 271

and abandoned land at the expense of cultivated

and ploughed areas; (iii) better farm management

practices.

†

only two sampling sites, and there were also

upward trends at two sites. The trends in TP can be

explained by changes in phosphorus emissions

from the municipalities.

†

and one statistically significant upward trend were

observed for the TN:TP ratio. The general decline

in the TN:TP ratios has promoted blue-green algae

blooms in the recipient, Lake Peipsi. Conse-

quently, it is imperative that decision makers and

managers focus much more attention on removal of

phosphorus.

†

differences in TN trends, even though the changes in

agricultural practices have been similar in the two

. 9. N:P ratio in the Emajogi river at the Joesuu site in 1987–

1.

countries. This may be partly due to differences in

the sizes of the river catchments and varying hydro-

geological conditions, although further studies are

needed to confirm that assumption.

Acknowledgements

We thank the Estonian Environment Information

Centre for providing the data. The study was

performed within the EC-funded project MANTRA-

East (Contract No. EVK1-CT-2000-00076). We are

also grateful to Patricia Odman for revision of the

English text.

References

Agriculture 2001, 2002. Yearbook. Statistical Office of Estonia.

Tallinn, 71 pp.

Agriculture in Estonia, 1996, 1997. Edited by M. Koov. Janeda

Training and Advisory Centre, Estonia, 97 pp.

Berankova, D., Ungerman, J., 1996. Nonpoint sources of pollution

in the Morava river basin. Water Science and Technology 33,

127–135.

Blinova, I., 2001. Riverine load into L. Peipsi. In: Lake Peipsi.

Meteorology, Hydrology, Hydrochemistry. Tartu, pp. 94–96.

Deelstra, J., Vagstad, N., Loigu, E., Vasiljev, A., Jansons, V., 1998.

Interactions between hydrology and nutrient runoff in small

agricultural catchments. A comparative study of Estonian,

Latvian and Norwegian catchments, 20th Nordic Hydrologic

Conference, vol. 1, pp. 120–129.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273272

Evans, R.O., Parsons, J.E., Stone, K., Wells, W.B., 1992. Water

table management on a watershed scale. Journal of Soil and

Water Conservation 47, 58–64.

Fertilizer Yearbook 2001, 2002, vol. 51. FAO Statistics Series, 268

pp.

Gambrell, R.P., Gilliam, J.W., Weed, S.B., 1975. Nitrogen losses

from soils of the North Carolina coastal plane. Journal of

Environmental Quality 4, 317–323.

Gilliam, J.W., Skaggs, R.W., 1986. Controlled agricultural drainage

to maintain water quality. Journal of Irrigation and Drainage

Engineering 112, 254–263.

Gilliom, R.J., Helsel, D.R., 1986. Estimation of distribution

parameters for censored trace level water quality data 1,

estimation techniques. Water Resources Research 22, 135–146.

Haraldsen, T.K., Loigu, E., Iital, A., Jansons, V., Vagstad, N., 2001.

Plant nutrients in soils and cereals in Norway and Baltic

countries. Jordforsk report no. 105/01. Norwegian Centre for

Soil and Environmental Research, 22 pp.

Hussian, M., Grimvall, A., Petersen, W., 2004. Estimation of the

human impact on nutrient loads carried by the Elbe River.

Environmental Monitoring and Assesment 96, 15–33.

Iital A., Loigu E., 2001. Agricultural runoff monitoring in Estonia.

In: Environmental Impact and Water Management in a

Catchment Area Perspective. Proceedings of the Symposium

dedicated to the 40th Anniversary of Institute of Environmental

Engineering at Tallinn Technical University. 24–26 September,

Tallinn, pp. 67–76.

Iital, A., Loigu, E., Vagstad, N., 2002. Nutrient losses and NP

balances in small agricultural watersheds in Estonia. In: XXII

Nordic Hydrological Conference, vol. 1. Røros, Norway 4–7

August, pp. 221–228.

Jaagus, J., Tarand, A., 1988. Sademete territoriaalne jaotus Eestis.

In: Eesti Geograafia Seltsi aastaraamat (in Estonian). Tallinn,

vol. 24, pp. 5–18.

Jarvekulg, A., 1993. Trophy of the water of Estonian rivers

and nutrients limiting the primary production. In: Water

Pollution and Quality in Estonia. Environment Data Centre.

National Board of Waters and the Environment, Helsinki, pp.

29–34.

Kauppi, L., 1979. Effect of drainage basin characteristics on the

diffuse load of phosphorus and nitrogen, Publications of the

Water Research Institute, National Board of Waters, Finland,

No. 30, pp. 21–41.

Keeney, D.R., DeLuca, T.H., 1993. Des Moines River nitrate in

relation to watershed agricultural practices: 1945 Versus 1980s.

Journal of Environmental Quality 22, 267–272.

Kutra, G., Aksomaitiene, R., Rackauskaite, A., 2002. Nitrogen

concentration in open watercourses as a result of leaching

from agricultural fields. Transactions of the Lithuanian

University and Lithuanian Institute of Water Management

18 (40), 13–19.

Libiseller, C., Grimvall, A., 2002. Performance of partial Mann

Kendall tests for trend detection in the presence of covariates.

Environmetrics 13, 71–84.

Loigu, E., 1993. Nutrient balance in surface water. In: Water

Pollution and Quality in Estonia. Environmental report 7,

Helsinki, pp. 18–22.

Loigu, E., Iital, A., 2000. Nutrient losses from two small

agricultural catchments in Estonia, Nordic Hydrological Con-

ference 2000, NHP-Report No. 46, pp. 139–145.

Loigu, E., Vassiljev, A., 1997. Evaluation of water quality

response to sudden changes in the amounts of fertilizers used

in Estonia. In: Hydrology an Environment. Proceedings of

the Baltic States Hydrology Conference. Kaunas, pp. 123–

130.

Loigu, E., Velner, H., 1985. Load and water quality in small rivers.

In: Ecological Modelling of Small Rivers and Water Bodies.

Proceedings of Soviet–Danish Symposium, June 17–19, 1981,

Denmark, Leningrad, pp. 57–60.

Mander, U., Kull, A., Tamm, V., Kuusemets, V., Karjus, R., 1998.

Impact of climatic fluctuations and land use change on runoff

and nutrient losses in rural landscapes. Landscape and Urban

Planning 41 (3/4), 229–238.

Mander, U., Kull, A., Kuusemets, V., Tamm, T., 2000. Nutrient run-

off dynamics in a rural catchment: influence of land-use

changes, climatic fluctuations and ecotechnological measures.

Ecological Engineering 14 (4), 405–417.

McMahon, P.B., Bohlke, J.K., 1996. Denitrification and mixing in a

stream-aquifer system: effects of nitrate loading to surface

water. Journal of Hydrology 186, 105–128.

Mourad, D.S.J., Van der Perk, M., Gooch G.D., Loigu E., Piirimae

K., Stalnacke P., 2003. GIS-based quantification of future

nutrient loads into Lake Peipsi/Chudskoe using qualitative

regional development scenarios. In: Proceedings of Diffuse

Pollution and Basin Management Conference (DipCon), Dublin,

17–22 Aug. 2003.10F GIS, pp. 105–111.

Noges, T., Blinova, I., Jastremski, V., Laugaste, R., Loigu, E.,

Skakalski, B., Tonno, I., 2002. Reduced nitrogen loading

enhance cyanobacterial blooms in Lake Peipsi. In: Sustainable

management of transboundary waters in Europe. Second

international conference, Miedzyzdroje, 21–24 April, 2002,

pp. 397–401.

Olah, J., Olah, M., 1996. Improving landscape nitrogen metabolism

in the Hungarian lowlands. Ambio 25, 331–335.

Pekarova, P., Pekar, J., 1996. The impact of land use on

stream water quality in Slovakia. Journal of Hydrology 180,

333–350.

Povilaitis, A., Tumas, R., Vagstad, N., 2003. Impact of agriculture

decline on nitrogen and phosphorus loads to Lithuanian Rivers.

Manuscript.

Prochazkova, L., Blazka, P., Kopacek, J., 1996. Impact of diffuse

pollution on water quality of the Vltava river (Slapy

Reservoir), Czech Republic. Water, Science and Technology

33, 145–152.

Rekolainen, S., 1989. Phosphorus and nitrogen load from forest and

agricultural areas in Finland. Aqua Fennica 199, 95–107.

Stalnacke, P., Tamminen, T., Vagstad, N., Wassmann, P.,

Loigu, E., Jansons, V., 1999. Nutrient runoff and transfer

from land and rivers to the Gulf of Riga. Hydrobiologia 410,

103–110.

Stalnacke, P., Grimvall, A., Libseller, C., Laznik, M., Kokorite, I.,

2003. Trends in nutrient concentrations in Latvian rivers and the

response to the dramatic change in agriculture. Journal of

Hydrology 283, 184–205.

A. Iital et al. / Journal of Hydrology 304 (2005) 261–273 273

Stalnacke, P., Vandsemb, S.M., Vassiljev, A., Grimvall, A.,

Jolankai, G., 2004. Changes in nutrient levels in some Eastern

European rivers in response to large-scale changes in agricul-

ture. Water Science and Technology 49, 30–36.

Tamm, T., 2001. Nitrate leaching from agricultural land in Estonia:

two case studies. In: First Baltic Symposium on Environmental

Chemistry. Abstracts. 26–29 September, 2001, Tartu, Estonia,

pp. 93–94.

Tamm, I., 2003. Adavere-Poltsamaa nitraaditundlikus piirkonnas

erinevate projektide raames kogutud andmete analuus uhtses

andmepangas veevarustuse parandamise voimaluste seisukohast

lahtuvalt. Too nr 3010 (in Estonian) AS Maves, Tallinn, 21 pp.

Tonderski, A., 1997. Control of nutrient fluxes in large river basins.

PhD Thesis. Linko ping Studies in Arts and Science No. 157,

Motala, Sweden, 50 pp.

Tumas, R., 2000. Evaluation and prediction of nonpoint pollution in

Lithuania. Ecological Engineering 14, 443–451.

Vagstad, N., Stalnacke, P., Andersen, H.-E., Deelstra, J.,

Gustafson, A., Ital, A., Jansons, V., Kyllmar, K., Loigu, E.,

Rekolainen, S., Tumas, R., Vuorenmaa, J., 2001. Nutrient losses

from agriculture in the Nordic and Baltic countries—results of

measurements in small agricultural catchments and national

agro-environmental statistics. TemaNord, 591 (74 pp).

Vassiljev, A., Loigu, E., Iital, A., 2001. Nutrient balance of the L.

Peipsi basin. Pollution load of L. Peipsi. In: Lake Peipsi.

Meteorology, Hydrology, Hydrochemistry, pp. 89–93.

Zablocki, Z., Pienkovski, P., 1999. The changes in mineral nitrogen

concentrations in stream and drain waters of Western Pomerania

in 1973–1994. In: Nitrogen Cycle and Balance in Polish

Agriculture. Institute for Land Reclamation and Grassland

Farming, Falenty, pp. 159–167.