Modeling the efficiency of drainage practices on acid sulphate soils at present and future climates at river Kyrönjoki, Finland

Transcription

1 Modeling the efficiency of drainage practices on acid sulphate soils at present and future climates at river Kyrönjoki, Finland - draft report Finnish Environmental Institute Maiju Kosunen

2 Abstract On the western coast of Finland streams are periodically acified and loaded with toxic metals as a consequence of leaching from acid sulphate soils (ASS). According to current knowledge, the leaching of acid water can best be abated by applying the following measures: control drainage (CD), lime filter drains (LFD) and controlled drainage system with additional pumping of water during dry periods (CPD). It has been noted that mostly the controlled drainage system can not keep the groundwater level high enough in order to prevent oxidation of ASS during dry seasons. Therefore pumping of additional water is needed. It has been observed that the acidity of discharge water starts to increase after heavy rains and the quality of water is most severe when a rainy autumn or heavy spring flood follows a dry summer. Therefore climate change, especially increased precipitation and amount of dry periods might affect the runoffs from AS soils in the future. The ionic flow model HAPSU was developed to simulate SO 4, H +, Fe, Ca²+ and Al leaching from the runoff areas build up with AS soils and non-acid soils in boreal conditions. The model takes into account boreal winter with freezing and melting as well as it contains the transport equations for heat, water, dissolved elements, oxygen and carbon dioxide in a soil column. The chemical part of the model includes reactions such as cation exchange, oxidation, reduction, precipitation, dissolution and weathering. Also water protection practices of controlled drainage and lime filter drains are included in the model. Here the HAPSU model was used for comparing the long-term efficiency of the different water protection measures of CD and LFD to discharge waters quality in the runoff area of Rintala by river Kyrönjoki. Also the effect of CPD was estimated roughly. The simulations were done for the reference period and scenario period The scenario simulations were done by utilizing the temperature and precipitation data calculated by the experts of SYKE for the period with the 19 climate scenario average. Also simulated groundwater level, evapotranspiration and discharges were viewed. Comparison between simulated and measured values was made and the sensitivity of the model recorded. Our simulation showed that the CD method had only a minor effect on the simulated ph of discharge water in both periods, and The efficiency was improved when the LFD method was used simultaneously with CD method and in the long-term simulations the effect fortified slightly. In the CPD groundwater level can be kept high also during dry periods by pumping extra water into the drains and the results probably will be even better. The simulated leaching of SO 4 from AS soil area showed the decreasing pattern in the 21 year period. Simulated values were in line with the measured ones, but simulated metal concentrations had a high variation. Based on earlier studies the HAPSU model needs further improvement especially in regard to the chemical processes and to the validity of the drainage practices. Model scenario runs give fresh perspectives into the future but when judging the model output, it has to keep in mind that model simulation might give unexpected results and therefore the uncertainty can be quite high. 1

3 Table of contents 1. Introduction What are AS soils and what makes them harmful? Description of AS soils in Finland Restoring AS soils Control drainage Lime filter drainage LFD + CD CPD Aim of the study HAPSU model Study area Kyrönjoki Rintala Baseline climate Precipitation Temperature Evapotranspiration PH variability and dependence on discharge Previous reviews PH and discharge measurements during Climate change and AS soils Climate scenarios Climate of the reference and scenario periods Possible effects of climate change on hydrology Groundwater level Discharges Simulated discharges for period Discharges of extreme years of the reference period Evapotranspiration Efficiency of water protection measures First HAPSU simulations Simulations with different water protection actions in reference and scenario periods Sulphate areas (SULFA) Whole (HAPSU) and side drainage basin (SIVU) areas Comparing groundwater simulations of the HAPSU- and Climate scenario models Evapotranspiration Extreme years of the reference period Corrected HAPSU simulations Simulations with different water protection actions in reference and scenario periods Sulphate areas (SULFA) Whole (HAPSU) area Comparing simulations of the HAPSU- and Climate scenario models Groundwater pool and groundwater level Evapotranspiration Discharge Comparing modeled Al, Fe, SO4 and ph values to measured values Sensitivity of HAPSU model and simulating the effect of subsurface pumping and installed plastics

4 8.1 Sensitivity Subsurface drainage level Lateral flow towards the drains Simulating the effect of plastic and pumping Still to review References Poster presentation (7th IASSC, Vaasa).72 3

5 1. Introduction 1.1 What are AS soils and what makes them harmful? Acid sulphate soils (ASS) are soils or sediments, which contain high amounts of sulphidic acids as their name assesses (Powell & Martens 2005). Acid sulphate soils are not necessarily harmful to the environment, but exposing them to oxygen might have severe causes (Talau 1999). These soils are usually exposed to air as a consequence of human practices, such as draining of agricultural fields (Bärlund et al. 2004). The artificial drainage of AS soils is still carried out in many places of the world, because the areas are turned into farmland, but also natural processes, such as the lifting of land as a consequence of ice age or changes in climate, may cause the AS soils expose to oxygen (Åström et al. 2007). If these sediments containing high amounts of sulphides are exposed to the air, iron sulphides oxidize and sulphidic acids are formed and as a result the ph of the soil might decrease under 4 (Talau 1999). Drainage exceeds the physical, chemical and biological processes that step by step ripen the soil (Palko 1994). Problems are also caused because in these conditions some metals, such as iron and aluminum become soluble in amounts which are toxic to the environment (Talau 1999). Eventually, discharge waters from AS soils are flowed to the rivers nearby. These freshwaters might be affected quite heavily of these discharges, since they usually contain low amounts of ions, which buffer against these harmful compounds and eventually discharges might result in kills of aquatic life (Talau 1999). The threshold ph value in which most of the ecosystem parts can survive is 5.5, and problems start to occur as the value decreases towards 5 (Huttu & Koskenniemi 1998). In the coasts of western Finland many streams are periodically acidified. PH might decrease down to 2.8, as a consequence of the leaching of hydrogen ions from AS soils and the amounts of toxic metals, which become soluble in these conditions, might also be quite high (Bärlund et al. 2004). 1.2 Description of AS soils in Finland The estimated areas of these soils in Finland have been rather diverse, since the definition of acid sulphate soils has varied between studies (Bärlund et al. 2004). The ph value of the soils is always used as a criterion, but the sulphur content is not always taken into account. Bärlund et al. have reviewed (2004) the different estimates on AS soils and it seems that depending on the criteria of acid sulphatic soils the estimates vary between ha (Yli-Halla et al 1999; Puustinen et al. 1994). Regardless of the study the limiting ph value has been at least 5 (Puustinen et al. 4

6 1994), many times even 3,5 (Yli-Halla et al. 1999). Typically the sulphidic sediments are covered with at least 0,5 meters of non-sulphidic materials (Joukainen & Yli-Halla 2003). Many of the international classification systems of AS soils examine the soil only until 150cm depth and therefore might not be well implemented in Finland where sulphidic material might oxidize even lower due to the occasionally low water table levels (Bärlund et al. 2004). This might cause surprises and undervalues when estimating the amounts of discharges from the AS soils. 1.3 Restoring AS soils When restoring the waters affected by runoffs from AS soils, the applied protection methods might strive for preventing the oxygenation. With knowledge of the present, leaching can best be decreased by applying countermeasures in the soils that cause acidity (Bärlund et al 2004). In the studies related to AS soils certain characteristics have been commonly measured and monitored. In addition to measuring ph values and metal concentrations, identifying the Redox (reductionoxygenation) potential of the soils is an important factor in AS studies (Bärlund et al. 2004; Powell et al 2005; Palko 1994). Water table level is an important factor when planning the remediation techniques for AS soils. The Redox - potential, which is positive in an oxidized profile, is measured to determine the chemical drainage depth of the soil (Palko 1994). Other important characteristics, such as soil texture and base saturation are also commonly determined (Bärlund et al. 2004). All these characteristics are important to determine, so that the planning and consequences of remediation can be evaluated. Quite a few water protection actions have been applied in order to enhance quality of river Kyrönjoki. Control drainage (CD), lime filter drainage (LFD) and a combination of CD and LFD are the most common ones recently used. Also other actions, such as surface liming, have been applied but not to such great extent in Finland. CD and LFD have had varying results at different sites. Also controlled subsurface pipe drainage (CPD) has been studied and implemented lately as a part of the CATERMASS project (Österholm et al. 2012, proceedings) Control drainage In control drainage (CD) the water table level of a cultivated field is tried to be kept as high as possible during the dry periods to enhance cultivation and to lower the table during spring floods to 5

7 improve drainage with the help of weir structures (Bärlund et al. 2004). Therefore it could be possible to stabilize the water management and to decrease the amounts of irrigation and runoff from fields by control drainage (Bärlund et al. 2005). During a 3-year monitoring program in Mustasaari, Åström et al. (2007) did not discover any notable impacts, except a lower concentration of Pb, in the water quality of the rivers after controlled drainage and overall noted that the technique did not raise the water table level as it was expected to. However, the CD technique has noted to work in some cases and is still applied and developed further Lime filter drainage In lime filter drainage (LFD), 5-7 % of lime (CaO) is mixed in the excavated soil material in a drain (Bärlund et al. 2004). The lime filter drains are usually constructed in connection with the normal subsurface drainage and the acid water has to pass through the lime filter material which surrounds the pipe drains (Weppling 1997). The LFD practice has seen to reduce acidity at sites where the sulphidic material lie deeper (2m, Ilmajoki), whereas at sites with sulphidic material closer to the soil surface (1m, Mustasaari) no such effect was recorded (Bärlund et al. 2005). So it seems that the results of these techniques are strongly related not only to the water table level but also to the drainage and exploitation history of the area. LFD can not be recommended for areas which are poorly leached, like Mustasaari, which has been drained only about 30 years ago (Bärlund et al. 2004) LFD + CD Also a combined method of lime filter drainage and control drainage (LFD+CD) has been applied as a remediation technique. In this practice both of these previously mentioned techniques are applied. The water table level is controlled with weir structures and the waters are attempted to neutralize with lime filters. The combination of CD and LFD has found to be quite successful. After a monitoring period it was noted (Bärlund et al. 2004) that in drainage waters Al concentrations decreased by %, ph values increased by 1-2 and critical 5,5 ph value was maintained at least for five years with this combination method. LFD and CD can possibly be recommended for intensively drained soils, where most of the acidity has already been leached to the rivers (Bärlund et al. 2005). 6

8 CPD Control drainage might not be effective enough to maintain the ground water level at the aimed level. Pumping of water is possibly needed during the dry summer months when evapotranspiration is high. CPD is a practice in which subsurface irrigation in addition to CD is applied. Also installation of plastic sheets in the soil to prevent lateral water flow has been added to this practice and this combined method has showed promising results in the test fields of Södenfjärden and Pedersjöre (Österholm et al. 2012, proceedings). 1.4 Aim of the study In this study the effect of control drainage and lime filter drains on drainage water ph and metal loads are estimated by utilizing the HAPSU model (Hutka et al. 1996). Also simulated groundwater level, evapotranspiration and discharges are viewed. Comparison between simulated and measured values is made and the sensitivity of the model recorded. Also the effect of subsurface pumping and installation of plastics in to the soil (to prevent lateral flow) is roughly simulated. In addition to thhe reference period ( ) the impact of changing climate is also reviewed for the period This was done by utilizing the temperature and precipitation data, simulated by the experts of SYKE for the period with the A1B average of 19 climate scenarios (GCMs) (Veijalainen et al. 2010: IPCC 2007) HAPSU model The ionic flow model HAPSU was developed in the 90's to simulate SO 4, H +, Fe, Ca²+ and Al leaching from the runoff areas build up with AS soils and non-acid soils in boreal conditions (Hutka et al. 1996). The model takes into account boreal winter with freezing and melting as well as it contains the transport equations for heat, water, dissolved elements, oxygen and carbon dioxide in a soil column. The chemical part of the model includes reactions such as cation exchange, oxidation, reduction, precipitation, dissolution and weathering. The model uses a variety of independent variables to explain for example the effect of soil type, climate and human practices on leaching. It consists of three parts: the main acid sulphate soil area (SULFA), the side drainage basin (SIVU) and the whole sulphate area (HAPSU), in which information from the SULFA and SIVU areas are combined. Controlled drainage and lime filter drains are included in the model. 7

9 1.4.2 Study area The study area Rintala (23km2), is located by river Kyrönjoki (Figure 1). Kyrönjoki is the biggest river in Southern Ostrobothnia and almost half of its fields (total ha) are in the area between Ilmajoki ( : ) and Ylistaro ( : ) (Teppo et al 2006). In the simulations we used a share of 50 % for acid sulphate soil and 50 % for the side catchment of Rintala. However, according to never studies the share of ASS area is estimated to be greater. Figure 1. Location of the study area Kyrönjoki The catchment area is 4920 km² (Ekholm 1993). Mean discharge for period at Kyrönjoki was 43 m³/ s (Saarinen et al. 2010). Kyrönjoki is often flooded and the acidity originating from the drained fields might be notable. The nature of river Kyrönjoki is still quite diverse (Teppo et al 2006). However, fish kills have been reported in the river already in 1834 and mass deaths in the 1970s' (Bärlund et al. 2004). 8

10 In a study of Saarinen et al. (2010) was noted that from nine large western Finnish coastal rivers the lake percentage (1, 2 %) was the lowest and farmland (25 %) and peat harvesting areas (1, 9 %) highest in the Kyrönjoki area (Saarinen et al. 2010). The area near river Kyrönjoki has gone through different kind of phases of exploitation and agricultural practices have been common for long. Drainage generalized widely in the 20 th century and between the Kyrönjoki catchments lost 23 lakes, mainly because of drainage (Hilden & Rapport 1993). Now the lake percentage at the area is very low, only 1, 2 % (Saarinen et al. 2010). Drainage and other anthropological disturbances have affected the quality of the river. Without acidification River Kyrönjoki would have become a eutrophied lowland river, such as many coastal rivers around the Baltic Sea, with high cyprinid fish populations (Hilden & Rapport 1993). The upper part of Kyrönjoki has also a high risk for floods. Embankments have been made in Rintala in to protect the area for floods (Teppo et al. 2006). Because of embankments and pump stations, the drainage at Rintala area has become more effective, which has again gives better conditions for the formation of harmful acid and metal loads (Teppo et al. 2006). However, near the Seinäjoki streambed Seinäsuu pump station prevents the groundwater level from decreasing as much as before during dry periods (Huttu & Koskenniemi 1998) Rintala The area of Rintala is situated northwest from the city of Seinäjoki and the coastline of river Kyrönjoki. The catchments nearby were also studied. The catchments that covered most of the study area were "area of Munakka" (42_031), "the upper part of river Kyrönjoki" (42_034) and "mouth part of river Seinäjoki" (42_071) (Figure 2.1). Also catchments called "area of Ilmajoki" (24_032) and "area of Tieksinluoma" are situated near Rintala. Most of the water quality measurement points were Sotaoja ( : ), mouth of Sotaoja ( : ), Seinäsuu pump station ( : ) and Seinäsuu upper part ( : ) (Figure 2.1). Since the soil type and cultivation history of the area might have an impact on leaching of ions and metals, the averages of the soil texture of the catchments, cultivated and slopes of the cultivated hills near Rintala area are presented (Figures 2.1, 2.2 and 2.3). 9

11 Figure 1.2 The catchments near Rintala area (area 42_031 continuing also to the West side of Sotaoja), river Kyrönjoki flowing in the middle. Soil texture at the closest catchments was mainly finer fine sand (0,02-0,06 mm, 49 %) (Figure 2.1). Barley (49 %), oat (14 %) and wheat (10 %) were the three most cultivated plants of the area (Figure 2.2). The fields were mainly not steep, while most (77%) of them had a slope of 0 0, 5% (Figure 2.3). Figure 2.1 Average of different soil textures of catchments (42_031, 071, 034, 032 and 039) (not weighted averages). 10

12 Figure 2.2 Average of cultivated crops of the catchments (42_031, 071, 034, 032 and 039). Figure 2.3 Average of slopes of the hills of the catchments (42_031, 071, 034, 032 and 039). 11

13 2. Baseline climate Climate variables measured by the stations close to Rintala area were utilized in the HAPSU simulations and review. Temperature and precipitation data were collected from the period The temperature data is originated from the measurements of the climate station in Kauhava ( : ). Some missing measurements were replaced with data from the Ähtäri station ( : ). The precipitation data was measured in Kauhava ( & ), Ylistaro (Pelma) ( : ) ( ) and Ilmajoki ( : ) ( ). 2.1 Precipitation The annual sums for precipitation (Figure 3.1) varied between 386 mm (year 1994) and 790 mm (year 1998). The average annual precipitation sum for the whole period was 528 mm. The summer was mainly the rainiest seasons, but also summers with low precipitation, such as 1999 and 2006 occurred during the reference period (Figures 3.2 & 3.3). Autumns with highest precipitation were in 1992 and 2004 (Figures 3.2 & 3.3). Figure 3.1 Annual precipitation sums of the Rintala area for the period

14 Figure 3.2. Precipitation sums for different seasons, , Ilmajoki, Hiiripelto (blue=winter, red=spring, green=summer, purple=autumn). Figure 3.3. Precipitation sums for different seasons, , Ilmajoki, Hiiripelto (blue=winter, red=spring, green=summer, purple=autumn). 13

15 2.2 Temperature Annual average temperatures varied between 2, 7 C (2010) and 5, 6 C (2000), whereas the average annual temperature for the whole reference period was 4, 02 C (Figure 3.4). Year 2010 was exceptional, having high summer temperatures but also low winter temperatures, which caused the average annual temperature (2,7 C) to be notably below the average of the reference period (Figure 3.4). Figure 3.4. Annual temperature averages of the Rintala area for the period Evapotranspiration The average annual evapotranspiration for months May - September at the Rintala area varied during the simulation period, standard deviation being 61,2 mm. Minimum occured in 1998 (389 mm) and maximum in 2006 (595 mm) (Figure 3.5). The annual average evapotranspiration for the whole period was 499 mm. 14

16 Figure 3.5. Annual average Class A evapotranspiration for months May-September measured from the Ylistaro, Pelma station ( : ) for the period PH variability and dependence on discharge 3.1 Previous reviews Climatic variables, such as precipitation, temperature and snow melt affect the recharge and discharge rates and therefore also the groundwater level. Hydrogen ions and metals flow away from AS soils in solid forms. If the discharge rate is higher, it is likely that more runoffs spread to the river. If discharge is low, the metals and H+ ions likely do not move as much as with higher discharge rates, keeping the ph rates of the river more neutral than with high discharges. Saarinen et al. (2010) noted a negative linear regression between the maximum discharge (log-transformed) and minimum ph for the autumn-winter flood period in the rivers which are severely affected by AS soils (Kyrönjoki and Lapuanjoki, R2 = 0,31 and 0,3, F = 22,03 and 20,31, p<0,001) and also with the moderately affected (Lestijoki, Kalajoki and Siikajoki; R2=0.49, 0.25 and 0.30, F=28.26, and 19.44, p<0.001). 15

17 Teppo et al. (2006) found the most notable change during a study period of to be that spring floods at river Kyrönjoki started to occur earlier, likely because of more temperate winters, also the acidity period was noted to start earlier in the 90s than before. So possibly because of earlier discharge peaks the acidification occurred earlier as well. 3.2 PH and discharge measurements during PH and discharge measurements taken at Kyrönjoki area during the period were collected from the Hertta database of Finnish Environmental Institute. First the values were changed in to hydrogen ion form (10^-(pH value)) and then the average was calculated from those values and afterwards changed back to ph form (-log(h+ value)). Mean is more commonly used when studying the ph variation, but here we used average values, so that comparison with the modeled average values would be possible. River Kyrönjoki has a notable and quickly changing variation is its discharge rates (Hutka et al. 2006). The acidity increases when moving down the stream and can be especially low (ph 5 or less) in many side rivers and ditches (Huttu & Koskenniemi 1998). This can be noted in the ph measurements from different areas of Kyrönjoki. In the middle of Sotaoja all of the monthly ph averages from period (Figure 4.1) were under 5,5 and the proportion of all measurements over 5,5 was only 21 %. In Seinäsuu pump station during period , the ph averages of July and August were over 5,5 and of all measurements 56 % were over ph 5,5 (Figure 4.2). In Skatila during period , most of the monthly ph averages were over 5,5 and of all measurements 72 % were over ph 5.5 (Figure 4.3). However, it has no be noted that the amounts of measurements varied between different months. There was for example only one measurement done in July and two measurements in February in Skatila during the period (Figure 4.3). However, there is an overall increasing trend in the ph values when moving further from the Rintala area (Figure 4.4) In Seinäsuu pump station during period the ph averages of February and July were over 5,5 and 57 % of the measurements were over 5,5 (Figure 4.5). In Skatila during period , most of the monthly ph averages were over 5,5 and 74 % of all measurements had ph over 5.5 (Figure 4.6). 16

18 Figure 4.1 Monthly average ph value and range for the period , measured from the middle of Sotaoja (121 measurements). Figure 4.2 Monthly average ph value and range for the period , measured from the Seinäsuu pump station (257 measurements). 17

19 Figure 4.3 Monthly average ph value and range for the period , measured from the Skatila ( : ) (173 measurements). Figure 4.4 Monthly average ph values for the period , measured from the middle of Sotaoja, Seinäsuu pump station and Skatila. 18

20 Figure 4.5 Monthly average ph value and range for the period , measured from the Seinäsuu pump station (482 measurements). Figure 4.6 Monthly average ph value and range for the period , measured from Skatila (418 measurements). 19

21 It might be difficult to see the variation in discharges when values are presented as annual averages. However, the variation between years can be noted and the greatest difference in annual discharge averages in Skatila for the period was m3/s, when the values of 2009 (9209 m³/s) was only 41 % of the value of 1992 (22537 m³/s) (Figure 5.1). There was also variation in the ph values measured in Skatila (Figure 5.2). It can be noted that discharge and ph values mostly act in opposite ways (Figure 5.3). However, when comparing the measured ph values to the measured discharge values, it has to be taken into account again that amounts of measurements differ. If the measurement dates are concentrated to a certain season, the average values might be misleading. All of the ph averages measured in Seinäsuu pump station, Sotaoja and Skatila are presented in Table 3. All of the stations did not have measurements every year. Figure 5.1. Annual discharge averages (419 measurements) measured in Skatila for the period

22 Figure 5.2. Annual ph values (419 measurements) measured in Skatila for the period Figure 5.3. Annual ph averages and discharge values for the period measured in Skatila. 21

23 Table 3. PH averages measured between Year Seinäsuu pumpst. Sotaoja Skatila Sotaoja middle , , ,8 4,4 5, ,2 4,6 6, ,4 4,6 5, ,7 4,3 5, ,6 5, ,2 5, ,7 5, ,8 6, ,7 5, ,9 5,6 4, ,2 6,0 4, ,9 5,3 4, ,8 5,2 4, ,0 5,7 4, ,9 5,4 4, ,8 5,3 4, ,7 5,7 4, ,1 5,9 4, ,1 5,5 4,9 4. Climate change and AS soils Based on the simulations done with the average of 19 global climate models, by the end of this century annual mean temperature and precipitation in Finland are estimated to increase by 3-7 C and 13-26%, largest increases occurring during winter time (IPCC 2007; Ruosteenoja & Jylhä 2007). For example the hydrological cycle might change since snowfall is predicted to decrease and summer and autumn rains to increase. The runoffs from AS soils are likely to change as well. The 22

24 acidity starts to increase easily after heavy rains and the worst case is when a rainy autumn or heavy spring flood follows a dry summer (Teppo et al. 2006). Therefore a changing climate might have a negative effect on the condition of the rivers near ASS. By using the data from climate scenarios done by Noora Veijalainen and Markus Huttunen from the Finnish Environmental Institute, possible effect of the changing climate on the AS soil runoffs could be simulated. 4.1 Climate scenarios Reference period for the climate scenario was and the data collected from the Herttadatabase of Finnish Environmental Institute from the measurement points of Nikkola, Hanhikoski and Skatila (Figure 6). The measured temperature and precipitation of the reference period were changed using the delta change approach, in which the monthly average temperature and precipitation changes from the scenarios are added to the measured temperatures and precipitations of the reference period (Veijalainen et al. 2010). The scenarios were calculated (by Huttunen & Veijalainen) as an average of 19 GCMs (IPCC 2007) by using the A1B emission scenario. The scenario used in the HAPSU simulations was made for the period , but also data for the period was received. Figure 6. Discharge- and other measurement stations used in the scenario (Nikkola, Hanhikoski and Skatila). Also Sotaoja shown in the map (1: ). 23

25 The received climate scenario data for hydrological changes in Rintala area was done with a hydrological watershed model in the Watershed Simulation and Forecasting System (WSFS; Vehviläinen et al. 2005) which is based on the Swedish HBW-model structure (Bergström 1976). The scenario was done by simulating the daily discharge and water level values (30 year period) with the WSFS in the reference and future time periods by utilizing the delta change approach and 19 climate scenarios (Veijalainen et al. 2010) Climate of the reference and scenario periods The average annual temperature of the Rintala area for the reference period was 3,09 C and 4,53 C for the period The average annual precipitation sums were 579 mm for period and 601 mm for the period On the HAPSU simulation period , years 2000 and 2010 were exceptional in terms of temperature and years 1994 and 1998 in precipitation. The measured annual average temperature for the year 2000 was exceptionally high (5,56 C) and low for the year 2010 (2,21 C). The measured annual precipitation sums were high for the year 1998 (790 mm) and low for the year 1994 (386 mm). The year 2010 was also unusual by having lower winter temperatures and higher summer temperatures than the simulated periods (Figure 7.1). The year 1998 also stands out by having very high summer precipitations (Figure 7.2). Therefore the climate- and hydrological conditions of these years might give an indication of what might be expected as a consequence of the changing climate. Figure 7.1. Monthly average temperatures for periods and (simulated) and years 2000 and 2010 measured near Rintala area. 24

26 Figure 7.2. Monthly average precipitation sums for periods and (simulated) and monthly precipitation sums for the years 1994 and 1998 measured near Rintala area. 4.3 Possible effects of climate change on hydrology Groundwater level The levels of groundwater will be affected by climate change as a consequence of shifted spring recharge towards winter and decreased summer recharge (Kundzewicz et al. 2007). Also the composition of vegetation and intensity of floods are likely affected, which again might have an impact on the groundwater level (Kundzewicz et al. 2007). However, impacts of climate change on groundwater level are dependent on the site characteristics and regional patterns. Okkonen et al. (2010) estimated that in northern Finland groundwater levels might enhance in winter because of the increased precipitation and decreased soil frost, whereas in summer increased temperatures and evapotranspiration might lower the groundwater levels. The highest summer soil moisture decreases will occur in northern Finland, whereas on southern Finland further reductions will probably not happen, since the minimum values of soil moisture are already so low during dry summers (Silander et al. 2006). So the effect of climate change on grounwater level varies even inside the area of Finland. The reagional patterns of present state can be estimated to move northeastwards and 25

27 generally groundwater levels are expected to increase during winters and decrease during summers (Silander et al. 2006) Discharges Climate change will likely affect discharges since temperature and rainfall are to increase. The earlier melting of snow and increased precipitation might enhance the winter discharges notably wheres during summer the discharges might decrease because of the increased evapotranspiration (Silander et al. 2006). On the other hand, climate change is said to increase extreme weather events, such as droughts, heavy precipitation and heat waves (IPCC 2007). As the probability of extremely dry as well as wet summers increase, the water levels and discharges might be higher or lower than at present time (Silander et al. 2006). The autumn-winter discharges (maximum discharge/mean discharge) were important factors to explain the acidity in rivers near AS soils of western Finnish rivers (Saarinen et al. 2010). Therefore if discharges are to change in the future, rivers near AS soils might be greatly affected Simulated discharges for period The simulated discharge values for the climate scenario period were also available in the water model system WSFS (KK625) of the Finnish Environmental Institute. As was said on the literature review, the autumn-winter discharges are predicted to increase in the future. This can also be seen in the simulated discharge values for Seinäsuu pump station and Sotaoja (Figures 8.1 and 8.2). The discharges are predicted to be higher in the autumn-winter months (October March) but during spring time the discharges are estimated to decrease (Figures 8.1 and 8.2). This is a consequence of higher precipitation during autumn-winter and lower and earlier snow melt at spring months. 26

28 Figure 8.1. Monthly discharge averages simulated for the Seinäsuu pumpstation for periods and Figure 8.2. Monthly discharge averages simulated for the Sotaoja for periods and

29 Discharges of extreme years of the reference period Discharges of some climatically extreme years of the reference period were viewed. This data was collected from the Hertta database and the values were measured, except the discharge values for Seinäsuu pump station were simulated and collected from the water model system WSFS (kk625). Compared to the averages of the reference period, year 1994 had a quite low average temperature (3,2 C) and low annual precipitation (386 mm), whereas year 2008 had a high average temperature (5,07 C) and annual precipitation (624 mm) (Table 2 and Figure 3.1). So year 2008 could roughly present a year of the estimated future climate. Variation can be seen for example when comparing figures 9.1 and 9.2 which are presenting years with different climate conditions at Seinäsuu pump station. During 1994, 31 % (11/35) of the measured ph values were under the critical level of 5, 5 (Figure 9.1). During 2008, as much as 75 % (15/20) of the measurements were under ph 5, 5 (Figure 9.2). As can be seen from Table 3, the average ph values for year 1994 was also notably higher (5, 4) than during 2008 (4, 7). The average discharge value was 0, m³/s for 1994 and 0, m³/s for For the measured values of Skatila the variation was not that notable. During 1994, 8, 7 % (2/23) of the measured ph values were under 5, 5 (Figure 9.3). During 2008, 6, 7 % (1/15) of the measured ph values were under 5, 5. So no clear difference was noted. Average ph value for 1994 was higher (5, 8) than for 2008 (5, 7) (Table 3). Average discharge value was 62, 8 m³/s for 1994 and 61, 6 m³/s for The ph and discharge values acted similarly in the HAPSU simulations (see Chapter 6.2.2). However, the amounts of measurements are quite low, so they do not really give that reliable information. 28

30 Figure 9.1. Measured ph and simulated discharge values (35) for the year 1994 (exceptionally low precipitation and average temperature) in Seinäsuu pumpstation. Figure 9.2. Measured ph and simulated discharge values (20) for the year 2008 (exceptionally high precipitation and average annual temperature), in Seinäsuu pumpstation. 29

31 Figure 9.3. Measured ph and discharge values (23) for the year 1994 (exceptionally low precipitation and average annual temperature) in Skatila. Figure 9.4. Measured ph and discharge values (15) for the year 2008 (exceptionally high precipitation and average annual temperature), in Skatila. 30

32 4.3.3 Evapotranspiration The increasing temperature is to raise evapotranspiration and lake evaporation in spring and early summer (Silander et al. 2006). Evapotranspiration and potential evapotranspiration are highest during summer, but the difference between evapotranspiration and potential evapotranspiration seems to increase in the future according to the climate simulations (Figure 10.1). However, the change in the evapotranspiration / potential evapotranspiration ratio is not that notable during summer as it is in winter and spring (Figure 10.2). This is explained when viewing at the temperature scenarios (Figure 7.1). In proportion the temperatures seems to rise most during winter and spring time. For example in December and January the average temperature increases between the reference period and scenario period by 1,91 and 2,15 C, whereas in June and July the increases are 1,2 and 0,98 C. Overall it seems that the evap. / potential evap. ratio will increase during winter and spring, but during dry summer months with high temperatures and a possible shortage in water, it might even decrease. From October to May evapotranspiration would increase more than the potential evapotranspiration and from June to September the increase in potential evapotranspiration would be greater than in evapotranspiration (Figure 10.1 & 10.2). Figure 10.1 Mothly evapotranspiration and potential evapotranspiration sums calculated as averages for the reference period and climate scenario

33 Figure 10.2 Mothly average temperature and ratio between evapotranspiration and potential evapotranspiration for the reference period and climate scenario Efficiency of water protection measures It is important to keep the groundwater level higher than the sulphidic level so that oxidation will not proceed (Bärlund et al. 2004). However, it seems that controlled drainage is not powerful enough to prevent further oxidation. At summer, i.e. higher evapotranspiration rates, the groundwater level is to decrease even though controlled drainage is applied and in the climate conditions of western Finland keeping the groundwater level high enough would require additional pumping of water into the drains (Bärlund et al. 2004). Figures 11.1, 11.2, 11.3 and 11.4 were done with the data of a previous study on acid sulphate soils of Ilmajoki (Bärlund et al. 2004). A monthly average (only for months 4-12, since there was not enough data for months 1-3) was calculated for the ph values of the areas, in which different water protection actions were implemented. It has to be noted that the amounts of measurements varied between months and years (year 1998 had 39 measurements, 1999:15, 2000:31 and 2001:14). Autumn to spring months had more measurements than the summer months (4:26, 5:22, 6:5, 7:1, 8:5, 9:7, 10:15, 11:21, 12:10). 32

34 The combination of control drainage and lime filter drainage showed to work most effectively in lowering the acidity of drainage waters (Figure 11.2). When no water protection measures were implemented (Figure 11.4), the average ph remained under 4, 5 during the whol period. Controlled drainage kept the ph higher than no water protection action, but the monthly average ph still remained under 5 (Figure 11.1). In the situation of no action and control drainage (Figures 11.1 and 11.4) it can be seen that acidification increases in autumn after groundwater level has lowered during dry summer months and possibly caused more sulphidic acids to form. The autumn precipitation has possibly released these acids from soil to the rivers. With lime filter drainage (Figure 11.3) and LFD + CD (Figure 11.2) there seems to be more variation in the ph values, but still the values decrease towards winter. Figure Average values (months April-December) for ph and ground water table level (GWT) in Ilmajoki , measured in at a plot in which CD was applied. 33

35 Figure Average values (months April-December) for ph and groundwater table level (GWT) in Ilmajoki , measured in at a plot in which CD+LFD was applied. Figure Average values (months April-December) for ph and groundwater table level (GWT) in Ilmajoki , measured in at a plot in which LFD was applied. 34

36 Figure 11.4 Average values (months April-December) for ph and ground water table level (GWT) in Ilmajoki , measured in at a plot in which no water protection methods were applied. 6. First HAPSU simulations The first simulations were done with the options of: no water protection actions (No action), control drainage at all sulphate area (CD), control drainage and lime filter drainage at 40 % of the sulphate area (CD_LFD 0,4) and control drainage and lime filter drainage at 70 % of the sulphate area (CD_LFD 0,7). The simulations were done for the period as well as for the climate scenario period The climate scenario is made for the perios , but for the HAPSU-simulations climate variables (temperature and precipitation) from only the first 21 years (7670 days) were utilized so the period would be consistent with the reference period ( ). The model was noticed to be sensitive for data containing climatical "extreme years". However, the simulations could be run by lowering the B coefficient of potential evapotranspiration (HAIB) from 0,16 to 0,04 in sulfa1.dat and sivu1.dat files. Also the share of applied lime filter drainage had to be lowered from 100% to 70 % in order to run the model completely. It was also noticed, that lowering of HAIB was too drastic, since it affected the evapotranspiration rates to an unrealistic extent. After 35

37 all, the percentage of 70 % for the coverage of LFD is far from reality and it seems that the model is not capable to run properly with such high values. Therefore more realistic simulations, with 5 % and 15 % coverage of LFD were done and presented in Chapter Simulations with different water protection actions in reference and scenario periods Sulphate areas (SULFA) The average ph values were increased notably when applying water protection actions (Figure 12.1). When applying control drainage in the whole sulphate area (SULFA) and lime filter drainage at 70 % of the sulphate area, the average ph value increases with almost one unit compared to a situation with no water protection actions. However, in all cases the average ph value seems to remain under the critical level of ph 5,5. Also it has to be noted that the increase of LFD from 40 % to 70 %, decreases the ph value in the scenario period (Figure 12.1). Average ph values seem to be higher in the scenario than period reference (Figure 12.1). This is unexpected, since the loads could be estimated to increase with changing climate. If temperature increases it could be estimated that the water table level would lower, which would again enhance the forming of sulphidic acids and with increased precipitation more metals and hydrogen ions would flow to the rivers. However, the increased amount of water would likely also reduce the concentration of hydrogen ions in solutes and therefore keep the ph value higher. As can be seen the average annual hydrogen ion load increases in the scenario period compared to period in all actions except CD+LFD 0,4 (Figure 12.2). Also the average annual water volume is to increase notably in the future (Figure 12.3), which partly explains the increase in ph values. The water volume seems to vary between different practices as well, being highest in control drainage (Figure 12.3). 36

38 Figure 12.1 Average annual water ph values with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. Figure 12.2 Annual average hydrogen ion load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. 37

39 Figure 12.3 Annual average water volume with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. Aluminum loads seem to be lower during the scenario period with No action. Only when applying CD and LFD at 70 % of the area, the loads in the are estimated to be greater in the future (Figure 12.4). When applying most water protection actions (CD+LFD 0,7), Al load seems to increase (Figure 12.4). However, sulphate loads seem to decrease gradually when increasing the intensity of water protection actions and even then the loads are estimated to be higher, except for No action, in the future climate than during the reference period (Figure 12.5). According to the simulations, iron loads are expected to be greater in the future climate than during the reference period and loads are decreased gradually when increasing the intensity of water protection actions (Figure 12.6). Runoff was unrealistically high in the scenario period for CD, which might partly a consequence of the changed HAIB coefficient (Table 4). 38

40 Figure 12.4 Annual average aluminum load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. Figure 12.5 Annual average sulphate load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. 39

41 Figure 12.6 Annual average iron load with different water protection practices simulated with HAPSU model for the baseline and climate scenario period in the sulphate area. Table 4. Simulation results for the sulphate area, annual averages, periods and No action CD CD + 0,4 LFD CD + 0,7 LFD Uncorrected precipitation, mm Evapotranspiration, mm Runoff, mm Water volume, m³ Hydrogen ions, mol ph 3,89 4,03 4,65 4,72 Aluminum, kg Sulphate, kg Iron, kg Uncorrected precipitation, mm Evapotranspiration, mm Runoff, mm Water volume, m³ Hydrogen ions, mol ph 3,95 4,16 4,87 4,73 Aluminum, kg Sulphate, kg Iron, kg

42 6.2.2 Whole (HAPSU) and side drainage basin (SIVU) areas When combining the simulation results from SULFA and SIVU areas, the ph values increase when adding the intensity of water protection actions, but still the annual averages of ph remain under 5,5 in each case (Figure 13.1). The monthly average ph values vary notably between different months, being lowest during spring and autumn (Figures 13.2 and 13.3). However according to the simulations, it seems that the variation in ph between different months is to increase in the future (Figure 13.3). For example in April the ph values appear to be lower in the future, whereas during summer time they might be higher. Still in each case the monthly average ph values seem to remain under the critical value of 5,5 (Figures 13.2 and 13.3). Figure 13.1 Average annual water ph values with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the whole area. 41

43 Figure 13.2 Monthly average water ph values with different water protection practices simulated with HAPSU for the baseline period in the whole area. Figure 13.3 Monthly average water ph values with different water protection practices simulated with HAPSU for the climate scenario period in the whole area. 42

44 The simulated ph and discharge values showed a similar kind of effect as was noted with the measured values in Chapter 4. The example years reviewed were 1994 and Year 1994 having low averages for annual precipitation and temperature and 2008 high averages for annual precipitation and temperature compared to the averages of the reference period. When discharge rates were lower, ph tended to increase (Figures 13.4 and 13.5). According to the simulations, the discharge was higher during year 2008 (average, m³/day) than in 1994(average, 8341 m³/day) (greater precipitation in 2008). The average annual ph was higher during 1994 (4,54 with no action) than in 2008 (4,42 with no action). Figure 13.4 Simulated daily discharge values for the side drainage basin area (SIVU) and ph values for the whole area (HAPSU) with the HAPSU model for the year

45 Figure 13.5 Simulated daily discharge values for the side drainage basin area (SIVU) and ph values for the whole area (HAPSU) with the HAPSU model for the year Water protection actions seemed to decrease the loads of aluminum, sulphate and iron in most cases, but the trends differed in the reference period and scenario period (Table 5). For example in the reference period, Fe loads decreased when increasing the amount of water protection actions. In the scenario period the trend was similar except for the peak when applying CD (Table 5). Table 5. Simulation results for the whole area, annual averages, periods and No action CD CD + 0,4 LFD CD + 0,7 LFD Hydrogen ions, mol Aluminum, kg Sulphate, kg Iron, kg Hydrogen ions, mol Aluminum, kg Sulphate, kg Iron, kg

46 6.2 Comparing groundwater simulations of the HAPSU- and Climate scenario models Evapotranspiration Values for evapotranspiration differed in the climate model and the first simulations of HAPSU model. Both predicted a rise in evapotranspiration, but with a presented a completely different kind of monthly trend (Figure 15.1). This was because of the change in the potential evapotranspiration coefficient (HAIB) and was corrected for the newer simulations (Chapter 7). The evapotranspiration seems to increase overall in both of the models, but in both cases the wintertime evapotranspiration is higher in the HAPSU simulations, whereas summer-time evapotranspiration is significantly higher in the climate scenario simulations (Figure 15.1). However, the evapotranspiration review only showed that the evapotranspiration coefficient had to be corrected in order to get realistic values for the summer months. Figure 15.1 Monthly average sums of evapotranspiration simulated with the HAPSU model (uncorrect parameters) for the sulphate area and simulated with the scenario model ( scenario), period

47 6.2.2 Extreme years of the reference period During the reference period also years 1994 and 2000 differed from the average. In 1994 the average precipitation and temperature was low, when in 2000 precipitation and temperature were both quite high. This could also be seen in the simulations for the groundwater and discharge. The change was noted more for discharge than for GWT. Simulated discharges with HAPSU model for the Rintala area and measured discharges for Hanhikoski showed notable variation between years 1994 and 2000 (Figures 16.1 & 16.2). So did the simulated discharge values with HAPSU and with the climate model for Seinäsuu pump station (Figures 16.3 & 16.4). During 2000 when precipitation was greater, also the discharge values were higher. However the HAPSU model seems to give a steadier evaluation for the discharges than the climate model (Figures & 16.4). This might partly be because the simulations are made for different areas and with different scales. Still the overall trends with the models are similar (Figure 16.3 & 16.4). Figure 16.1 Simulated discharges with HAPSU model for the Rintala area and reference data of the climate scenario measured in Hanhikoski for year

48 Figure 16.2 Simulated discharges with HAPSU model for the Rintala area and reference data of the climate scenario measured in Hanhikoski for year Figure 16.3 Simulated discharges with HAPSU model for the Rintala area and simulated discharges with the climate model (1) for Seinäsuu pump station for year

49 Figure 16.4 Simulated discharges with HAPSU model for the Rintala area and simulated discharges with the climate model (1) for Seinäsuu pump station for year Corrected HAPSU simulations The second simulations were done with the options of: no water protection actions (No action), control drainage at all sulphate area (CD), control drainage and lime filter drainage at 5 % of the sulphate area (CD_LFD5%) and control drainage and lime filter drainage at 15 % of the sulphate area (CD_LFD15%). The simulations were done for the period as well as for the climate scenario period The climate scenario is made for the perios , but for the HAPSU-simulations climate variables (temperature and precipitation) from only the first 21 years (7670 days) were utilized so the period would be consistent with the reference period ( ). 48

50 Input data from previous simulations in the Kyrönjoki area (Ilmajoki) were used in the simulations. Only temperature and precipitation data were changed according to the measured and simulated periods. The SULFA and SIVU areas were both 11,5 km² by size. Total area (HAPSU) was 23 km². The model was noticed to be sensitive for data containing high variation in weather events (ie. High pre., high/low T). However, the simulations were able to be run by lowering the B coefficient of potential evapotranspiration (HAIB) from 0,16 to 0,12 in sulfa1.dat and sivu1.dat files. This decrease was so mild that is was not noted to affect the evapotranspiration values. Some problems were found in the water balance of the model, but were corrected later (see Chapter 8). However, the correction did not seem to specially affect the amounts of ph and metal loads, so those results are presented here. 7.1 Simulations with different water protection actions in reference and scenario periods Sulphate areas (SULFA) As was noted before, the combined method of CD and LFD has worked in increasing the ph of drainage waters. This was also seen in the HAPSU simulation runs. According to the simulations, the ph values seem to increase slightly in the future (Figure 18.1). Lime filter drainage increased the ph value notably, but applying LFD in a larger area (15%) did not increase the value further. This is possibly related to the characteristics of the model. It would be expected that applying this water protection action at a larger sulphate soil area, would result in an increased ph value. It seems that after increasing the parameter of LFD over 10% the output of hydrogen ions in discharge water remains constant. PH is presented on a logarithmic scale, therefore the effects of water protection actions are more clear when viewing at the hydrogen ion loads (Figure 18.2). The water volume is much greater in the climate scenario period, which might also affect the ph loads by diluting the hydrogen ion concentration (Figure 18.3). 49

51 Figure 18.1 Average water ph values with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. Figure 18.2 Annual average hydrogen ion load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. 50

52 Figure 18.3 Annual average water volume with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. All of the water protection actions seem to decrease the annual aluminum loads (Figure 18.4). Again has to be noted that increasing the area of LFD from 5% to 15% does not decrease the Al loads (Figure 18.4). Also, the loads seem to be smaller in the scenario period (Figure 18.4). HAPSU model does not simulate notable effects of water protection actions on the amount of sulphate loads, but the loads seem to be greater in the future climate (Figure 18.5). The effect of protection actions on iron loads is similarly quite low, but contrary to the aluminum loads, the quantity seems to increase in the scenario period compared to the reference period (Figure 18.6). 51

53 Figure 18.4 Annual average aluminum load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. Figure 18.5 Annual average sulphate load with different water protection practices simulated with HAPSU for the baseline and climate scenario period in the sulphate area. 52

54 Figure 18.6 Annual average iron load with different water protection practices simulated with HAPSU model for the baseline and climate scenario period in the sulphate area. Runoff seems to be much greater, when applying water protection actions (Table 6). This is because control drainage shifts the water table level up. However, the increase seems to be more significant in the scenario period than in the reference period. This is not only explained with the increased precipitation. For example the CD option gives unrealistic values for the scenario period by predicting that the sum of runoff and evapotranspiration is much higher than precipitation. Therefore the water balance of the model has to be fixed (Chapter 8). The highest ph values seemed to focus on the summer months (Figure 18.7). This was also seen in the previous chapters with the measured values. One explanation is that the discharge is lower because of the dry summer months, and therefore ph is higher. When applying lime filter drainage with control drainage, the ph values increase at least with one unit every month (Figure 18.7). 53

55 Table 6. Results of the simulations for the sulphate area, annual averages for periods and No action CD CD + 5%LFD CD + 15%LFD Uncorrected precipitation, mm Evapotranspiration, mm Runoff, mm Water volume, m³ Hydrogen ions, mol ph 3,76 3,89 4,59 4,57 Aluminum, kg Sulphate, kg Iron, kg Uncorrected precipitation, mm Evapotranspiration, mm Runoff, mm Water volume, m³ Hydrogen ions, mol ph 3,84 4,03 4,84 4,77 Aluminum, kg Sulphate, kg Iron, kg Figure 18.7 Monthly average water ph values with different water protection practices simulated with HAPSU for the baseline period in the sulphate area. 54

56 The groundwater values given in the "sulfa1.lis" file, were transformed to actual groundwater level values with the formula (4-x)*(-1). This way, when reviewing single years, the groundwater level might occasionally reach the soil surface (Figure 18.8). But when looking at the monthly averages and their minimums and maximums (min and max of months averages, not daily), it can be seen that variation is highest during summer months (Figure 18.9). There was also a lot of variation in the precipitations of those months, which affects the results, since the model seems to be sensitive to the precipitation values. The groundwater level is high during spring after the snow has melted and lowest during summer months, when the evaporation is high and rises again as a cause of the increased autumn precipitations (Figure 18.9). Figure 18.8 Groundwater level simulated with HAPSU model with different water protection actions for year 2000 in the sulphate area. 55

57 Figure 18.9 Monthly average and variation in groundwater depth when applying control drainage Whole (HAPSU) area In the whole study area, the difference in ph values between water protection actions was not as notable as with the sulphate area (Figures 19.1 and 19.2). The ph values for the whole area are calculated as an average of the sulphate area and SIVU area (which do not change when applying different actions). Figure 19.1 Monthly average water ph values with different water protection practices simulated with HAPSU for the baseline period in the whole area. 56

58 Figure 19.2 Monthly average water ph values with different water protection practices simulated with HAPSU for the climate scenario period in the whole area. Also in the whole study area, the simulations showed notable decreases in the hydrogen ion loads when applying lime filter drainage (Table 6). Sulphate and iron loads seemed to increase in the future, but aluminum decreased. CD seemed to increase the hydrogen ions load in the scenario period, but then again decrease after adding of LFD (Table 6). Table 6. Results of the simulations for the whole area, annual averages for periods and No action CD CD + 5%LFD CD + 15%LFD Hydrogen ions, mol Aluminum, kg Sulphate, kg Iron, kg Hydrogen ions, mol Aluminum, kg Sulphate, kg Iron, kg

59 7.2 Comparing simulations of the HAPSU- and Climate scenario models Some variables simulated with the climate model and HAPSU model were compared to note possible differences or similarities between models. These variables were viewed as examples and also within certain years, especially years with extreme climates. Compared to other years of the reference period, year 1998 was high with precipitation and low with temperature, year 1994 was low with both and year 2000 high with both Groundwater pool and groundwater level The groundwater pool and groundwater level can not directly be compared, but they still should act with a same kind of trend. Compared to the climate model, HAPSU model seemed to react to the high summer evaporations with a delay (Figure 20.1). Responses to precipitation or evapotranspiration extremes was different between models. Figure 20.1 Ground water table level (simulated with HAPSU) and groundwater pool (simulated with climate model 1) for years Evapotranspiration The evapotranspiration was very similar with the simulations of HAPSU model and climate model after again increasing the HAIB coefficient (Figures 21.1 and 21.2). However, climate model showed no evapotranspiration during winter months, when HAPSU model shoved some on average. Both models shoved and increase in the summer evapotranspiration values. This increase is likely to happen if the summer temperatures increase in the future. 58

60 Figure 21.1 Monthly averages of evapotranspiration simulated with the HAPSU model for the sulphate area and measured for Nikkola upper part (climate scenario reference data ), period Figure 21.2 Monthly averages of evapotranspiration simulated with the HAPSU model for the sulphate area and simulated with the scenario model ( scenario), period Discharge The differences in discharge values were hard to interpret, since the values are mainly quite low. Models did not seem to have similar kind of peaks at the same time. Climate model, showed the spring peak well, but HAPSU did not (Figures 22.3 and 22.4). Instead HAPSU showed higher peaks in the late summer, same time in 1994 and This is surprising, since years 1994 and 1998 were 59

61 totally different in precipitation. These factors must be viewed further and discharges can maybe be looked at shorter periods so that the smaller differences can be seen. Figure 22.3 Simulated discharges with HAPSU model for the Rintala area and simulated discharges with the climate model 1 for Seinäsuu pump station for year Figure 22.4 Simulated discharges with HAPSU model for the Rintala area and simulated discharges with the climate model 1 for Seinäsuu pump station for year

62 7.3 Comparing modeled Al, Fe, SO4 and ph values to measured values Several zero-values occur in the output data of the sulphate area Therefore median and minimum concentrations were not calculated for the sulphate area. However, modeled concentrations for the whole Rintala area (Table 7.4) are similar to the measured values from Kyrönjoki (Table 7.1 and 7.2). Overall, the concentrations measured in Ylipää bridge (Table 7.2) are lower and the ph is higher compared to the ones measured in Seinäsuu pump station and Skatila (table 7.1). The concentrations for Al and SO4 measured in Kyrönjoki seem to be higher than the background values for whole Finland (Table 7.3). The measured Fe concentrations and ph values fall in to the normal category. Also the modeled Al and SO4 concentrations with no action (Table 7.4) are higher and Fe concentrations and ph values similar to the background values. Table 7.1 Backround values from the Hertta-database of Syke, measured in Skatila and Seinäsuu pump station. Skatila variable period N average median min max Al (mg/l) ,68 1,60 0,27 4,36 Fe (mg/l) ,66 1,60 0,18 3,00 SO4(mg/l) ,20 89 Seinäsuu pump station ph ,93 5,8 4,1 7,2 Table 7.2 Backround values measured in Kyrönjoki, Ylipää bridge (Vuorenmaa et al. 2010). variable N median min max Al(mg/l) 12 1,0-1,8 Fe(mg/l) 12 1,65-2,6 ph 12 6,3 6,1 - Table 7.3 Typical (90%) concentrations in Finnish streams (Lahermo et al. 1996). variable values normally average median Al(mg/l) 0,02-0,25 0,134 0,095 Fe(mg/l) 0,06-2,6 0,91 0,68 SO4(mg/l) ,7 3,5 ph 4,7-6,6 5,91 61

63 Table 7.4 Simulated HAPSU values for the whole Rintala area average mediaani min max average mediaani min max NoAction Al(mg/l) 1,4 0,2 0,0 19,5 1,2 0,3 0,0 19,6 Fe(mg/l) 1,8 0,7 0,1 10,8 2,1 0,9 0,1 9,8 SO4(mg/l) 48,5 1,5 0,5 640,1 44,7 7,8 0,5 647,3 ph 4,3 5,0 3,7 5,3 4,3 4,9 3,8 5,3 CD Al(mg/l) 0,9 0,2 0,0 19,5 0,64 0,17 0,02 19,62 Fe(mg/l) 1,5 0,6 0,1 8,5 1,78 0,67 0,14 11,09 SO4(mg/l) 30,4 1,3 0,5 640,1 24,43 1,38 0,52 647,30 ph 4,3 5,0 3,8 5,3 4,3 5,0 3,80 5,26 CD_LFD5% Al(mg/l) 0,8 0,2 0,0 21,1 0,53 0,14 0,00 21,83 Fe(mg/l) 1,4 0,6 0,1 9,0 1,75 0,66 0,12 10,93 SO4(mg/l) 30,1 1,2 0,5 696,8 25,23 1,39 0,52 719,20 ph 4,8 5,1 4,1 6,8 4,9 5,1 4,16 7,02 8. Sensitivity of HAPSU model and simulating the effect of subsurface pumping and installed plastics 8.1 Sensitivity The sensitivity of the model was tested by changing some parameters in sulfa.dat files. Already the simulations, where LFD was set to a high level (70%, in chapter 6), revealed for example that HAIB is an essential factor and it should be lowered carefully. Also high levels of LFD (>10%) prevent the model from working, unless HAIB is lowered. Also the effects of changing parameters "subsurface drain level" and "lateral flow towards the drains" in sulfa1.dat were tested Subsurface drainage level The subsurface drainage level has to be set twice for each year. First it was set to 1.2 in 123 rd day and 0,5 in the 130 th day. 62

64 In the second simulations it was set to 1.2 in 130 th and 240 th days (Also the ZSO(m) value in sulfa1.dat was changed to 0,5. It gave unrealistic values so was set back to 1.2). This increased the variation in metal loads between different actions. Since the GWT value has to be corrected ((4- x)*(-1)), was noticed that the option of 1.2 and 0,5 simulates the differences in GWT better. This way "control drainage" seems to keep the groundwater level higher than "No Action", which is realistic. So the 1.2 in 123 rd day and 0.5 in the 130 th day was used in the main simulations Lateral flow towards the drains The effect of the factor KUIA, lateral flow towards the drains, was also tested. This value has to be set for each month. At first the values were: 0.75, 0.75, 1.5, 1.5, 0.75, 0.5, 0.5, 0.5, 0.75, 0.75, 0.75 and These values were multiplied by 0 and the simulations for sulfa were made. Then same was done by multiplying the values by 0.1, 0.25, 0.5 and then by The groundwater levels and metal values were viewed. When KUIA is zero, no water flows out of the system and groundwater level is near soil surface (Figure 23.1). Increasing these values even slightly lowers the groundwater level quite fast (GWT, NA, O -> GWT, NA, 0.25). If control drainage is applied, water table level is of course even higher (Figure 23.2). When KUIA is zero, evapotranspiration was noted to be lower than with other options. But when KUIA is more than zero, also evapotranspiration increases. However, further increase in KUIA, does not affect the evapotranspiration any more. KUIA affected the GWT notably, but the effect on metal loads was not so significant, at least when viewing the whole area (Figure 23.3). Figure Effect of KUIA on groundwater level, No Action. 63

65 Figure Effect of KUIA on groundwater level, CD. Figure Effect of KUIA on aluminum loads. It was also noticed that if KUIA is decreased severily, the ratio between Al and Fe is more realistic, at least with control drainage. When KUIA is increased, the load of iron increases, but aluminum stays almost the same. Surprisingly the water content of soil (in sulfa2.lis) was not affected by the decrease of KUI. Changing KUI does not seem to affect the water balance values severily. For example multiplying the values for each month by 0,25 decreased runoff by about 20% (Table 8). 64

66 Table 8. Simulation results and differences for normal KUI values and 0,25 x KUI values. 1 0,25 diff(%) 1 0,25 diff(%) precipitation ,00 mm evapotranspiration ,97 mm runoff ,79 mm water in total ,66 m3 hydrogen ions 1,073 0,922 0,86 mol/m ,86 mol Al 45,64 45,58 1,00 g/m ,00 kg SO4 1872, ,72 0,95 g/m ,95 kg Fe 65,529 46,475 0,71 g/m ,71 kg 8. 2 Simulating the effect of plastic and pumping During dry summer months subsurface pumping water and installing plastics in the soil have been shown to work in keeping the groundwater level higher (Österholm et al. 2012, proceedings). We also tried to simulate this with HAPSU. This can be roughly done either by decreasing the KUIA values or by increasing precipitation. First we tested this by increasing precipitation by 0,5 mm each day during months 5-9. This increase seemed too drastic, since the runoff values were increased by over 100%, but evapotranspiration was not affected almost at all. Aluminum and sulphate loads were decreased, but iron was increased. This was expected since iron reduces (Fe3+ -> Fe2+) into a more mobile form when water content rises. First the precipitation was increased by 0,2mm/day during periods , every year. This addition did not affect the runoff notably. We tested different variations to see what would be the best option to simulate pumping. Options were: added precipitation + decreased KUI, added pre + original KUI, normal precipitation + decreased KUI, normal precipitation + original KUI. The decrease in KUI was 10% and increase in precipitation 24mm/year. With control drainage, the affect was seen more clearly when increasing precipitation than decreasing KUI (Figure 24.1). With increased precipitation, groundwater level is shifted during the dry periods and water does not drop under 1,2m. Therefore sulphidic layers at Rintala would remain under water. 65

67 Figure 24.1Groundwater levels with different KUI and precipitation values. Compared to the option of original KUI and original precipitation, increased pre + orig.kui lifted GWT by 50mm/year, increased pre + 0,9KUI by 57mm/year and orig.pre + 0,9KUI only by 1,5 mm/year. Runoff is greatest with the option of added precipitation and original KUI and lowest with original precipitation and 0,9KUI (Table 9). There are no great changes in the metal concentrations, likely because the changes in precipitation are so small. Table 9. Metal and precipitation with different KUI and precipitation values. orig_pre+orig.kui added_pre+orig.kui added_pre+0,9kui orig_pre+0,9kui precipitation, mm/year evapotranspiration, mm/year runoff, mm/year H+, mol/m2 1,11 1,07 1,05 1,07 Al, g/m2 49,1 45,6 46,4 49,6 S04, g/m2 1943,6 1872,1 1867,2 1918,2 Fe, g/m2 68,1 65,5 60,1 60,9 66

68 9. Still to review It seems that simulating the effect of pumping either by reducing KUI parameters or adding precipitation has an effect on the metal and hydrogen ion runoffs as a consequence of the rise in groundwater level. However, precipitation can not be increased greatly, so that the water balance will not be affected too much. Also, the groundwater level is lifted more than precipitation is added, since the evaporation does not increase as much as water volume increases. The effects of changed parameters on other variables also have to be reviewed. Also possible changing of other parameters (evapotranspiration coefficients parameters affecting the lateral flow towards the drains, parameters describing dependence between water content and matric potential) or changing KUI parameters only for summer months (when pumping occurs) has to be considered. The used method is quite rough, and it is possible that the sufficient simulation of pumping and plastics would require a coding instead of changing parameters. In Finnish conditions there could be more dry summers and wet autumns in the future. Also mobilization of metals from acid sulphate soils could increase. Acidity peaks from AS soils occur after dry summers and heavy autumn rains (Teppo et al. 2006). The model is quite sensitive for changes in precipitation, so it is likely that these peaks could be simulated. Long periods should be reviewed more instead of averages. Loads of Al (Figure 25.1) and SO4 (Figure 25.2) decrease during the 21-year simulation period. Load of Fe is rather stabile (Figure 25.1). It has to be studied whether the differences between water protection practices and simulation periods can be seen in the development of these loads. 67

69 Figure 25.1 Al and Fe loads during whole simulation period with CD. Figure SO4 loads during whole simulation period with CD. 68

70 References Bergström, S Development and application of a conceptual runoff model for Scandinavian Catchments. Report RHO no 7. Swedish Meteorological and Hydrological Institute, Norrköping Bärlund, I. Tattari, S., Yli-Halla, M. & Åström, M Effects of sophisticated drainage techniques on groundwater level and drainage water quality on acid sulphate soils Final report of HAPSU project. The Finnish Environmental Institute. Report: 732 Bärlund, I. Tattari, S., Yli-Halla, M. & Åström, M Measured and simulated effects of sophisticated drainage techniques on groundwater level and runoff hydrochemistry in areas of boreal acid sulphate soils. Agricultural and Food Science. 14: Ekholm, M Suomen vesistöalueet. Vesi- ja ympäristöhallinnon julkaisuja A126, Helsinki. 166 s. Hilden, M. & Rapport, D Four centuries of cumulative impacts on a Finnish river and its estuary: en ecosystem health-approach. Journal of Aquatic Ecosystem Health 2: Hutka, R., Laitinen, T., Holmberg, M., Maunula, M. & Schultz, T Happamien sulfaattimaiden ioninvirtausmalli (HAPSU). Suomen ympäristö 8. ISSN Huttu, U. & Koskenniemi, E. Vaasa Rintalan pengerryksen valumavesien happamuuden vähentäminen Kyrönjoella. Alueelliset ympäristöjulkaisut: 69 Intergovernmental Panel on Climate change 2007: the physical science basis summary for policy makers. Contribution of working group I to the fourth assessment report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge Joukainen, S. & Yli-Halla, M Environmental impacts and acid loads from deep sulfidic layers of two welldrained acid sulfate soils in western Finland. Agriculture, Ecosystems and Environment 93: Kundzewicz, Z.W., Mata, L.J., Arnell, N.W., Döll, P., Kabat, P., Jiménez, B., Miller, K.A., Oki, T., Sen, Z. & Shiklomanov. I.A Freshwater resources and their management. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson, Eds., Cambridge University Press, Cambridge, UK, Lahermo, P., Väänänen, P, Tarvainen, T. & Salminen, R Suomen Geokemian Atlas, osa 3: Ympäristögeokemia purovedet ja sedimentit. Geologian tutkimuskeskus, Espoo. ISBN

71 Okkonen, J., Jyrkama, M. & Kløve B A conceptual approach for assessing the impact of climate change on groundwater and related surface waters in cold regions. Hydrogeology journal 18: Palko, J Acid sulphate soils and their agricultural and environmental problems in Finland. Academic Dissertation. Acta Univ. Oul. C 75 Puustinen, M. Merilä, E., Palko, J. Seuna, P., Kuivatustila, viljelykäytäntö ja vesistökuormitukseen vaikuttavat ominaisuudet Suomen pelloilla. National Board of Waters and Environment, Research Report A 198. Powell, B. & Martens, M. A. A review of acid sulfate soil impacts, actions and policies that impact on water quality in Great Barrier Reef catchments, including a case study on remediation at East Trinity Marine Pollution Bulletin 51: Ruosteenoja K, Jylhä K Temperature and precipitation projections for Finland based on climate models employed in the IPCC 4th assessment report. In: Proceeding of the third international conference on climate and water, 3 6 September Helsinki, Finland, p Saarinen, T., Vuori, K.M., Alasaarela, E. & Kløve, B Long-term trends and variation of acidity, CODMn and colour in coastal rivers of Western Finland in relation to climate and hydrology. Science of the Total Environment 408 (2010) Silander, J., Vehviläinen, B., Niemi, J., Arosilta, A., Dubrovin, T., Jormola, J., Keskisarja, V., Keto, A., Lepistö, A., Mäkinen, R., Ollila, M., Pajula, H., Pitkänen, H., Sammalkorpi, I., Suomalainen, M. & Veijalainen, N Climate change adaptation for hydrology and water resources. FINADAPT Working Paper 6, Finnish Environment Institute Mimeographs 335, Finnish Environment Institute, Helsinki Teppo, A., Tolonen, M., Korsu, K., Sivil, M., Koivurinta, M., Marjomäki, T., Koivisto, A.M., Latvala, J. & Rautio, M.R Kyrönjoen yläosan vesistötöiden vaikutus ja Kyrönjoen tila vuosina Suomen Ympäristö 18, Länsi-Suomen Ympäristökeskus Talau, M.J Acid Sulfate Soil Management Priority Areas in the Lower Clarence Floodplain. Report. Department of Land and Water Conservation, Sydney Vehviläinen B, Huttunen M & Huttunen I Hydrological forecasting and real time monitoring in Finland: the watershed simulation and forecasting system (WSFS). In: Innovation, advances and implementation of flood forecasting technology, conference papers, October 2005, Tromsø, Norway Veijalainen, N., Dubrovin, T., Marttunen, M. & Vehviläinen, B Climate Change Impacts on Water Resources and Lake Regulation in the Vuoksi Watershed in Finland. Water Resour Manage 24:

72 Vuorenmaa, A., Vuori, K-M., Siimes, K. & Mannio, J Happamien sulfaattimaiden vesistövaikutusten seuranta vuonna Suomen Ympäristökeskus, , MaaMet. Weppling, K On the assessment on feasible liming strategies for acid sulphate waters in Finland Tartu University Press. 196 pages. Dissertationes Geographicae Universitatis Tartulensis. Yli-Halla, M. Puustinen, M. & Koskiaho, J Area of acid sulphate soils in Finland. Soils Use and Management 15: Åström, M. Österholm, P. Bärlund, I. & Tattari, S Hydrochemical Effects of Surface Liming, Controlled Drainage and Lime-Filter Drainage on Boreal Acid Sulfate Soils. Water Air Soil Pollut 179: Osterholm, P., Astrom, M. & Sundstrom, R. 2005: Assessment of aquatic pollution, remedial measures and juridical oblications of an acid sulphate soil area in western Finland. Agricultural and food science, vol 14:44-56 Österholm, P., Virtanen, S., Uusi-Kämppä, J., Rosendahl, R., Westberg, V., Mäensivu, M., Ylivainio, K., Yli-Halla, M & Turtola, E Minimizing sulfide oxidation on acid sulfate farmlands by enhanced controlled drainage and subsurface irrigation. Geological Survey of Finland, Guide th International Acid Sulphate Soil Conference, Vaasa, Finland. Proceedings volume. Eds., Österholm, P., Yli-Halla, M & Eden, P. 71

73 Maiju Kosunen 1, Seija Virtanen 2, Sirkka Tattari 1, 1 Finnish Environment Institute SYKE, P.O. Box 140, FI Helsinki Finland ²Department of Food and Environmental Sciences, P.O. Box 27, FI University of Helsinki, Finland Modeling the efficiency of drainage practices at present and future climate scenarios on acid sulphate soils in Finland Background Drainage practices such as controlled drainage, lime filter drains and controlled drainage system with additional pumping of water during dry periods have been suggested to abate acid sulphate soil problems. In this study, the long-term effectiveness of these techniques is studied with the HAPSU model using future climate scenario data. Moreover, new experimental data is used to test the validity of the model. Ionic Flow Model for Acid Sulphate (AS) Soils The ionic flow model HAPSU was developed to simulate SO 4, H +, Ca 2+, Fe and Al leaching from the runoff areas build up with AS soils and non-acid soils in boreal conditions (Hutka et al. 1996). The model consists of normal drainage practice, lime filter drains and controlled drainage (Bärlund et al. 2004). Also the pumping of additional water into the drains can be simulated with the HAPSU model. Water protection practices and climate scenarios The simulations were done using the Kyrönjoki temperature and precipitation data in the period The climate simulations ( ) were done by utilizing the temperature and precipitation data (average of 19 GCMs) from the available A1B climate scenario for the period (Veijalainen et al. 2010). Efficiency of the following water protection practices were simulated with HAPSU: Control drainage (CD) Lime Filter Drainage (LFD) Pumping of additional water (CPD) Weir structures installed in the cultivated field 5-7 % of lime (CaO) mixed in the excavated soil material in a drain Lysimeter experiment Plastics installed in the soil and extra water pumped during summer months in addition to CD In this study the model performance was also tested by the data from a lysimeter experiment of AS soil. In the lysimeter experiment AS soil monoliths (Sulfic Cryaquepts), which included oxidized sulfuric B horizons and a reduced sulfidic C horizon, were subjected to two different water management treatments, normal drainage and waterlogging of soil. The response of soil ph on the different treatments was analyzed from pore water during the experimental period and in the soil profile before and after the experiment. The chemical part of Hapsu model was verified against that data. Figure 4. Simulated SO 4 and Fe concentrations for period (21 years) for the whole Rintala area (mixed water). µg/l date Fe, µg/l SO 4, mg/l mg/l 700 The simulated leaching of SO 4 from AS soil area applied with CD showed a decreasing trend during the 21 year period, whereas it settled on a constant value approximately after 10 years from start. The simulated leaching of Fe remained rather constant SUBMODEL 1 SUBMODEL 2 Acid sulphate soils Non-acid soils water leaching quality water SUBMODEL 3 Mixed water delay period load quality leaching quality exchange sites cations minerals: cations, micronutriens Output variables leaching SO 2-, H +, Ca 2+, Fe, Al 3+ 4 soil solution: water oxygen cations anions metals Figure 2. Shematic presentation of the HAPSU model. sulphides: FeS, FeS 2, MnS metals precipitates, complex forms: cations, anions, metals cation exchange weatheringof minerals oxidation/reduction reactions precipitation/dissolution reactions, build-up of complexes transport Table 1. Measured and modeled values for metal concentrations and ph for whole Finland and Kyrönjoki. Typical concentrations for Finnish streams (Geochemical Atlas of Finland 3 (Louhela et al.1996)) Measured average (Kyrönjoki: Skatila for Al, Fe, SO 4, Seinäsuu pump station for ph) Modeled average with Hapsu for the whole area (No action, period ) Modeled average with Hapsu for the whole area (CPD_added prec., period ) Modeled average with Hapsu for the whole area (CD + LFD5%, period ) Modeled average with Hapsu for the whole area (No action, period ) Modeled average with Hapsu for the whole area (CD + LFD5%, period ) Al(mg/l) Fe(mg/l) SO 4 (mg/l) ph ± 0.9 ( , n=205) 1.7 ± 0.7 ( , n=60) 1.4 ± ± 2 38 ± 17.9 ( , n=119) 48.5 ± ± 0.9 ( , n=482) 4.3 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.4 Concentrations for Al and SO 4 measured in the Kyrönjoki area are higher than the background values for whole Finland. Modeled parameters show a similar pattern as the measured ones. High standard deviations are seen in the modeled values, while the variation is lower in the measured values. The ph values were increased and concentrations of Fe, Al and SO 4 decreased when the CPD or CD + LFD methods were applied. river Kyrönjoki 11,5km km Figure 1. Location of the study areas. Grounwater table Discharge 11,5km 2 Water protection practices Normal drainage Evaportanspiration Irrigation Capillary rise Percolation river Seinäjoki Rintala soil Lysimeter experiment Ap B2g Bgjc BCgc Cg slightly acid ASS Helsinki soil ASS Soil surface Horizons Drainage level Vaasa km Waterlogging Irrigation Evaportanspiration Capillary rise Groundwater flow Seinäjoki Helsinki Figure 3. Acid sulphate soils submodel was tested on the ph data from a lysimeter experiment in which AS soil monoliths (Sulfic Cryaquepts) were drained or waterlogged. BCgc Figure 5. The ph of soil horizons in the beginning and at the end of the experiment and the results of simulation. Depth (m) a Normal drainage ph Ap B2g Bgjc BCgc Cg b Depth (m) Ap B2g Bgjc Cg FINLAND Groundwater table Discharge In normal drainage (a) the acidity of AS soil horizons increased in the simulation alike in the lysimeter experiment. However, in the simulation the decrease of ph was more intense in the Cg horizon and milder in the B horizons than in the lysimeters. Waterlogging (b) caused the rise of ph in the lysimeter experiment, but in the simulation only a slight rise of ph was observed in the B horizons. The ph of the Cg horizon remained about neutral in the simulation alike in the experiment. Reason for the slighter response of ph on waterlogging in the simulation compared to the experiment was supposedly that HAPSU takes into account only the chemical reduction of Fe and ignores the microbial catalyst. weir Waterlogging ph Ap B2g Bgjc BCgc Cg The ph values in the beginning The end / Lysimeter experiment The end / HAPSU simulation Conclusions and future challenges References HAPSU model shows the differences of water protection practices on metal and hydrogen ion loads. There is a rise in ph values and decrease in concentrations of SO 4 and Al after adding the water protection practices of CD, CPD and LFD to the model (Table 1.). Also the simulated metal concentrations are similar to the measured values. Climate change is likely to affect the runoffs from AS soils. The acidity starts to increase easily after heavy rains and the worst case is when a rainy autumn or heavy spring flood follows a dry summer. To examine the effect in a climate with even higher temperatures and precipitation values further simulations could be done with HAPSU for the period The lysimeter experiment provided evidence for the importance of microbial catalyst in redox processes. Because the chemical part of HAPSU takes into account only the chemical reduction of Fe and ignores the microbial catalyst, this part of model needs further development. Based on earlier studies the HAPSU model needs further improvement especially in regard to the chemical processes and to the validity of the drainage practices. Model scenario runs give fresh perspectives into the future but when judging the model output, it has to keep in mind that model simulation might give unexpected results and therefore the uncertainty can be quite high. Bärlund, I., Tattari, S., Yli-Halla, M. & Åström, M Effects of sophisticated drainage techniques on groundwater level and drainage water quality on acid sulphate soils Final report of HAPSU project. Finnish Environment: 732, 68p. Hutka, R., Laitinen, T., Holmberg, M., Maunula, M. & Schultz, T Happamien sulfaattimaiden ioninvirtausmalli HAPSU (Abstract in English: Ionic flow model for acid sulphate soils HAPSU) Suomen ympäristö 8, 154 p. Veijalainen, N., Dubrovin, T., Marttunen, M. & Vehviläinen, B Climate Change Impacts on Water Resources and Lake Regulation in the Vuoksi Watershed in Finland. Water Resource Management 24: Layout: Erika Várkonyi / SYKE, 8/2012