Results and confidence

Fulfilment of MAI in 2016 and progress since the reference period (1997-2003)

According to the revised HELCOM nutrient reduction scheme adopted in the 2013 HELCOM Ministerial Declaration (HELCOM 2013a) reduction requirements were set for nitrogen inputs to the Baltic Proper, Gulf of Finland and Kattegat and for phosphorus inputs to Baltic Proper, Gulf of Finland and Gulf of Riga.

The Kattegat is the only sub-basin out of three with set reduction targets for nitrogen inputs where the input was significantly below MAI in 2016 (Results figure 1 and Results table 1a). However, since the reference period (1997-2003), statistically significant reduction of nitrogen input has been achieved to all sub-basins except the Gulf of Riga, where no reductions has been observed (Results figure 2). The highest observed input reduction was to Kattegat (24%) and Danish Straits (23%), the lowest to the Bothnian Bay (8%) and Gulf of Riga (0%), respectively (Results figure 1). Nitrogen inputs to the entire Baltic Sea has been reduced on 16% since the reference period.


Results table 1. The trend-based estimate for normalized annual inputs of (a) nitrogen and (b) phosphorus during 2016.

The table also contains data on statistical uncertainty, the remaining reduction needed to reach MAI and inputs in 2016 including statistical uncertainty in percentages of MAI. Classification of achieving MAI is given in colours: green=MAI fulfilled, yellow= fulfilment is not determined due to statistical uncertainty, and red=MAI not fulfilled. (Units in columns 2-5: tonnes per year).NOTE: For consistency with MAI no rounding (to tenth, hundreds or thousands) has been performed in the indicator.

Table 1a.

Table 1b.

*As adopted by the 2013 HELCOM Copenhagen Ministerial Meeting (HELCOM 2013a).

**Exceedance of MAI is caused by statistical data uncertainty.


None of the 3 sub-basins, the Baltic Proper, Gulf of Finland and Gulf of Riga, for which reduction targets for phosphorus inputs were set, fulfilled the requirements in 2016 (Results figure 1 and Results table 1b). However, a certain reduction of phosphorus inputs was observed in all sub-basins. Statistically significant reduction was achieved in Kattegat, Baltic Proper and Gulf of Finland. The highest input reduction since the reference period (1997-2003) of 71% was achieved to the Gulf of Finland (Results figure 1). Phosphorus inputs to the entire Baltic Sea have been reduced by 25% since the reference period mainly due to reductions to Gulf of Finland and Baltic proper.

Compared to the first evaluation of MAI fulfilment (Svendsen et al. 2015), in 2016 EMEP revised the modelled nitrogen air deposition to the Baltic Sea for 1995-2012. This resulted in an increase of the annual deposition to the Baltic Sea of 16 to 23%. The increase of annual nitrogen deposition to the individual sub-basins was between 9 and 27 %. Then, the EMEP deposition model has once again been updated since the previous assessment (HELCOM, 2018). The update mainly concerned changing of the deposition grid resolution from 50km*50km to 0.1°*0.1° (approx. 11km*11km grid). The higher resolution was applied for computation of the annual nitrogen deposition 2000-2016. The update of the model resulted in revision of the previously reported deposition data. Modelled annual nitrogen deposition was reduced to Bothnian Bay (between 10-18%) and GUR (4-11%) and elevated to Gulf of Finland (1-10%), Danish Straits (6-12%) and Kattegat (1-4%). Though, the changes of airborne input to the entire Baltic Sea were less than 1%.

Changes in inputs of total nitrogen and phosphorus since the reference period are estimated based on the average normalized inputs in the reference period and inputs in 2016 resulting from the analysis of long term trends (1995-2016). 

Results figure 1. Reductions of annual inputs of nitrogen (left) and phosphorus (right) achieved in 2016 since the reference period 1997-2003 (in %). The annual inputs in 2016 and in the reference period were calculated using normalized annual data. The arrows indicate decreasing (↓) inputs, while the colours indicate if the change was statistically significant.



Normalization is used for the annual riverine and atmospheric inputs to reduce the impact of inter-annual variations of the inputs caused by weather conditions (primarily variations in precipitation). With normalisation the comparability of the inter-annual inputs increases, facilitating trend detection and also identification of effects of undertaken measures in the catchment areas. Without normalization, the effects could be disguised by large natural annual variation of precipitation and river flow.

Trend analyses show statistically significant steady reduction of total inputs of nitrogen to the Baltic Sea amounting to 20 % from 1995 to 2016 (Results figure 2). Correspondingly, from 1995 to 2016 total nitrogen inputs also show a steady significant decrease for Kattegat (29 %). For the remaining basins besides the Gulf of Riga, break points were identified when evaluating the trends of total nitrogen inputs, dividing the time series in two sections. Baltic Proper has a significant decrease with 16% from 1995 to 2000, and with 11% from 2000 to 2016. A significant nitrogen input increase of 28 % was observed in the Gulf of Finland during the 1995 to 2004 period, and a 26 % decrease from 2004 to 2016. Significant reductions were also calculated for Danish Straits during 1995-2003 (21%) and from 2003 to 2016 (15 %). For the Bothnian Bay a significant increase took place until 2007 (10%). A significant reduction of nitrogen input took place in the Bothnian Sea after 2003 (21%). No changes of input to the Gulf of Riga were determined during the whole period of 1995-2016.

Trends for total phosphorus inputs to the Baltic Sea reveal steady statistically significant reduction of 30% from 1995 to 2016 (Results figure 2). Correspondingly, from 1995 to 2016 phosphorus inputs also show steady significant decrease to Kattegat (19%) and to Baltic Proper (25 %). Break points were detected when evaluating the trend in the time series for the remaining five sub-basins. For four sub-basins there are significant decreases in total phosphorus inputs in the first section of the time series:  Bothnian Bay 18% (1995 to 2002), Bothnian Sea 19% (1995 to 2002), Danish Straits 31% (1995 to 2000), and for Gulf of Riga an increase with 26% (1995 to 2007). Significant decrease of phosphorus inputs after break points in time series are determined for Bothnian Sea (5 % 2002-2016), Gulf of Finland (69% 2009-2016) and Gulf of Riga (16% 2007-2016). The marked and abrupt reduction in total phosphorus inputs from 2009 to 2016 to Gulf of Finland is probably connected to rapid changes of inputs due to measures on point sources in the Russian catchment. There is large inter-annual variation due to a) uncertain data prior approximately 2009 prevented efficient flow normalization and b) an abrupt reduction of point source inputs. This results in very high uncertainty in the total phosphorus inputs to the Gulf of Finland (54 % according to Results table 1b). Also for Gulf of Riga there is higher uncertainty on estimated phosphorus inputs in 2016 (9 %) compared with the remaining 5 sub-basins (4-5%). For nitrogen estimated 2016 inputs the uncertainty are also highest to Gulf of Finland (7%) the inputs to the remaining sub-basins was between 2 and 4% (according to Results table 1a).

Results figure 2. Actual total air- and waterborne annual input of nitrogen (TN) and phosphorus (TP) to the Baltic Sea and sub-basins from 1995 to 2016 (tonnes). The normalized annual inputs of nitrogen and phosphorus are given as a black line. The trend line for normalized total nitrogen and phosphorus input is given as a grey line with markers. In cases when a break point divides the trend to two parts, the second part (called trend 2) is shown by a green line without marker. (Solid trend line shows statistically significant trend and dotted line - not statistically significant trend). The MAI as adopted by the 2013 HELCOM Copenhagen Ministerial Meeting (HELCOM 2013a) is shown as the bold dotted blue line.


Total nutrient input to the Baltic Sea varies significantly depending on wet or dry weather conditions. For example, 2010 was a very wet year in the southern part of the Baltic Sea catchment area, hence the actual (non-normalized) nutrient inputs were very high to e.g. Baltic Proper (Results figure 2) and relatively high to the whole Baltic Sea. Additionally, atmospheric deposition was also rather high in 2010.

Airborne nitrogen deposition on the Baltic Sea is a major nitrogen source, and therefore the input of nitrogen has been divided in airborne and waterborne inputs, to evaluate changes in these major pathways separately. Airborne input of nitrogen to the Baltic Sea in 2016 shows statistically significant reduction of 29% since 1995 (Results table 2). This reduction slightly varies between sub-basins from 26% to the Gulf of Finland up to 32% to Kattegat. In comparison to airborne deposition, waterborne input of nitrogen to the Baltic Sea has been significantly reduced by about 13% since 1995. This reduction varies widely between sub-basins. Statistically significant reduction of waterborne input was found for Danish Straights (41%), Kattegat (27%), Bothnian Sea (12%) and Baltic Proper (12%). Reduction of waterborne input to other sub-basins is not statistically significant including the 16% reduction to the Gulf of Finland. Waterborne input of nitrogen to the Bothnian Bay and Gulf of Riga has slightly increased by 2016. But these changes are also within statistical deviation.


Results table 2. Difference in annual average normalized air- and waterborne inputs of nitrogen in 1995 and 2016 in percent.

Sub-basinChange in airborne N inputs since 1995 (%)Change in waterborne N inputs since 1995 (%)
Bothnian Bay-284,9
Bothnian Sea-29-12
Baltic Proper-29-12
Gulf of Finland-26-16
Gulf of Riga-273,8
Danish Straits-28-41
Total Baltic Sea-29-13



Actual airborne and waterborne inputs in 2016

In 2015, the average water flow was about the same as average for the year 1995-2015 (1% lower). Flow lower than average was observed to the Gulf of Riga (33%) and to the Baltic Proper (30%). On the other hand, the flow to the Kattegat (9%), Danish Straits (22%) and Bothnian Bay (30%) was comparatively higher than average for the considered period. The total input of nitrogen was about 773,000 tonnes, and the portion of atmospheric deposition was about 30%. The total phosphorus input to the Baltic Sea in 2014 was about 24,000 tonnes with a contribution of atmospheric deposition by about 9% (Results table 3 and Results figure 3).


Results table 3. Annual average water flow as well as actual annual waterborne and airborne inputs of phosphorus and nitrogen to the Baltic Sea sub-basins in 2016. Average flow 1995-2016 is shown for comparison.

Average flow 1995-2016 (m3/s)
Flow 2016 (m3/s)
Nitrogen 2016 (t)
Phosphorus 2016 (t) ​ ​
Bothnian Bay345437195355662505980627831812964
Bothnian Sea2937237136818201005691814993941893
Baltic Proper347727342294191085803379999314104610360
Gulf of Finland350538011000731214011221340251504175
Gulf of Riga1078959809507330882802156932250
Danish Straits22422232460252805774012831051388


Results figure 3.  The total actual inputs of water- and airborne nitrogen (left) and phosphorus (right) to the Baltic Sea in 2016.


Confidence of the indicator status evaluation

The confidence is affected by the certainty of the quality of the nutrient input data, the trend in the inputs and the uncertainty of MAI, in relation to how far the nutrient inputs are from MAI:

The assessment is based on continuously improving national data and assessment methods. The quality of national data on input of nutrients has been significantly improved in the last years since introduction of the WEB reporting application and four stages of data quality control, which's made datasets more consistent. Nonetheless, data quality differs between sub-basins, which is reflected in the overall statistical uncertainty of the input computation which vary from few percentages in Kattegat to more than 50% in the Gulf of Finland. The uncertainty of input computation is taken into account in the assessment of fulfilment of the maximum allowable input of nutrients (MAI). The statistical uncertainty of MAI identification is generally low. But limitations in the model used to calculate MAI cause moderate uncertainty for nitrogen MAI to Danish Straits and phosphorus MAI to Bothnian Sea. The statistical confidence on the assessment of fulfilling MAI for basins with reductions requirement can be considered as high.