Assessment of nutrient retention in Hungarian rivers, based on long term water quality monitoring data

Background: Nutrient retention in rivers is a key part of diffuse nutrient pollution emission estimation processes, as model performance can only be validated at monitoring points on rivers. In-stream retention can be modelled by complicated water quality models like QUAL, but in many cases only empirical relationships provide the means of estimating the retention e.g. in MONERIS model. This latter model have been used in Hungary as well, and the current work analyse themonitoring data of the national water quality monitoring network in order to check the model’s performance on estimating in-stream nutrient retention. Data series length differ substantially at different monitoring stations, narrowing down the time gap in which retention calculation is possible. All monitoring stations that have at least monthly sampling have been included in the study, but only a fraction of them have more frequent measurements.

Methodology: This study follows a very simple methodology.

Step 1. Monitoring stations have been selected based on the following criteria: • neighbour stations, no tributary rivers, small relative diffuse load• neighbour stations with tributary between them, also monitored

Step 2. Monitoring station data statistics have been calculated• Number of data per calendar year• Average value for calendar year• Average value for x year periods• Standard deviation for calendar year

Step 3. Retention calculation for years, where the number of data is greater or equal than 10 both on upper and lower stations

• Load based retention calculation if data were sufficient

,where Lds and Lus are upstream and downstream nutrient loads respectively

• Concentration based retention calculation if Q was not known

,where Cds and Cus are upstream and downstream nutrient loads respectively

Step 4. Results were compared to Moneris equations used for retention estimation (Behrendt et al. 2000). Yearly average water temperature was approximated with 11 C° for Monerisretention estimation

𝑅𝑒𝑡𝑙𝑜𝑎𝑑 = 1 −𝐿𝑑𝑠

𝐿𝑢𝑠

𝑅𝑒𝑡𝑐𝑜𝑛𝑐 = 1 −𝐶𝑑𝑠

𝐶𝑢𝑠

Zala: • Frequently monitored catchment (~1600 km2)

• small river• Large q increase between

stations• Unknown tributary loads• Only „soft” estimation is

possible based on concentration values

Remarks• Significant river retention in

the 80’s• + conc. gradient after 2000• Likely cause: river bed inner

loading after major reduction in upstream point loads

0

100

200

300

400

500

600

1960 1970 1980 1990 2000 2010 2020

PO

4-P

MIK

KR

OG

/L

PO4-P concentrations Zala

Zalabér Zalaapáti

-0.5

0

0.5

1

1.5

0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Ret

enti

on

Hidraulic load (m/y)

TP retention

1994-2006 2007-2013 MONERIS equation

Danube

Tisza

Drava

Maros

-0.2

-0.1

0

0.1

0.2

0.3

0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Ret

enti

on

Hidraulic load (m/y)

DIN retention

1976-2005 2006-2013 MONERIS equation

Danube

Tisza

Drava

Hernád

Results

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0 2000 4000 6000 8000 10000 12000 14000 16000 18000Ret

enti

on

(-)

Hidraulic load (m/y)

TN retention

2007-2013 MONERIS equation 1977-1993

Danube

Tisza

Drava

Maros

Conclusions• None of the three nutrient parameters examined give back the empirical relationship, used in MONERIS

model, accurately• At higher hydraulic loads the retention values are close to zero, therefore fit well to the values calculated

by the Moneris equation• At lower hydraulic loads the average retention values between sampling points show larger variation,

including negative retention in many cases• Retention is changing along the Danube and Tisza, while there is also a temporal change in river nutrient

retention in both rivers (Fig. R5, R6)

Possible causes of unclear results

• Both TP and TN measurements and Q measurement are inaccurate therefore retention calculations are not accurate either (PO4-P and DIN gives a chance for control)

• Monthly sampling used as minimum criteria is still not accurate• Sampling at neighbour monitoring stations are sometimes delayed by many days, also causing inaccuracy

in the retention values• At smaller rivers, the increase of the flow is significant between monitoring points, but the loads from

the incoming flows are generally unknown, i.e. the smaller the river the more uncertain is the calculation

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

1985 1990 1995 2000 2005 2010 2015PO

4-P

RET

ENTI

ON

Zalabér - Zalaapáti

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

1985 1990 1995 2000 2005 2010

RET

ENTI

ON

PO4-P retention (Conc.) RSD

Szigethalom - Ráckeve Ráckeve - Tass

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

1985 1990 1995 2000 2005 2010

RET

ENTI

ON

PO4-P retention (Conc.) Danube

Fajsz-Baja Dunaföldvár - Fajsz Lineáris (Fajsz-Baja) Lineáris (Dunaföldvár - Fajsz)

R² = 0.5051

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1992 1997 2002 2007 2012 2017RET

ENTI

ON

TP retention (Conc.) Tisza

Upper Tisza Mid Tisza Lineáris (Upper Tisza) Lineáris (Mid Tisza)

R² = 0.2473

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

1985 1990 1995 2000 2005 2010 2015

TN R

ETEN

TIO

N

TN retention (Conc.) Zala

TN Lineáris (TN)

Fig R1,2,3: Retention values against Hydraulic load for TN, TP and DIN fractions

Fig R4,5,6: Retention values against time at different locations of different rivers

Zsolt JolankaiBudapest University of Technology and Economics, Department of Sanitary and Environmental Engineering

Műegyetem rkp. 3., H-1111 Budapest, Hungary (E-mail: jzsolt@vkkt.bme.hu)

Zalabér

Zalaapáti

River Basins 2017 - International Conference on Monitoring, Modelling and Management of River Basins, 19-20, June, 2017, Vienna

Fig R7,8,9: Temporal variations of in-stream conc. and ret. in Zala river