Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology CHALMERS UNIVERSITY OF TECHNOLOGY Master’s Thesis ACEX30 Gothenburg, Sweden 2020 Dimensioning Groundwater levels A case study of the TK Geo method Master’s thesis in the Master’s Programme Infrastructure and Environmental Engineering IDA NORDBERG MASTER’S THESIS ACEX30 Dimensioning Groundwater Levels A Case study of the TK Geo method Master’s Thesis in the Master’s Programme Infrastructure and Environmental Engineering IDA NORDBERG Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology CHALMERS UNIVERSITY OF TECHNOLOGY Göteborg, Sweden 2020 I Dimensioning Groundwater Levels A Case study of the TK Geo method Master’s Thesis in the Master’s Programme Infrastructure and Environmental Engineering IDA NORDBERG © IDA NORDBERG, 2020 Examensarbete ACEX30 Institutionen för arkitektur och samhällsbyggnadsteknik Chalmers tekniska högskola, 2020 Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology Chalmers University of Technology SE-412 96 Göteborg Sweden Telephone: + 46 (0)31-772 1000 Department of Architecture and Civil Engineering Göteborg, Sweden, 2020 I Dimensioning Groundwater Levels A Case study of the TK Geo method Master’s thesis in the Master’s Programme Infrastructure and Environmental Engineering IDA NORDBERG Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology Chalmers University of Technology ABSTRACT Dimensioning groundwater level is a term that is used within the area of geo- construction and infrastructure projects. The level is used to calculate the water pressure that an underground construction should resist. The concept of dimensioning groundwater levels is implemented in Eurocode and governing documents from the Swedish Transport Administration (Trafikverket) and it implies a similar treatment as other loads to the construction in terms of safety. Groundwater levels fluctuate resulting in difficulties deciding on a single value as a dimensioning groundwater level. It is not only the evaluation method that arise questions within the industry. It has also arisen questions of who is responsible for these assessments and how should the levels be dealt with in large infrastructure projects where the whole landscape is affected, and the hydrogeological conditions are changed fundamentally? There is no simple answer to these questions but in an attempt to contribute to the industry, this master thesis aims to perform a case study where one common evaluation method for dimensioning groundwater levels is tested and compared to a statistical analysis. The master’s thesis project also includes a literature study that describes the fundamentals about groundwater level fluctuation. The literature study also gives an overview of how loads and groundwater levels are treated according to Eurocode and Swedish regulations. The tested evaluation method for dimensioning groundwater levels in the case study is the method presented by the Swedish Transport Administration in their governing document, TK Geo. The method is developed by Chester Svensson and Göran Sällfors in the 1980s and is an easy way to establish dimensioning groundwater levels for construction sites mainly. The method is notated the TK Geo method in this report. The TK Geo method uses measurement series from a reference well to estimate groundwater levels at a site. In this thesis project, the TK Geo method is used to estimate the dimensioning groundwater levels at two case study sites. The results are then compared to statistical analysis of measurement series from the case study sites. The results show that the TK Geo method can predict dimensioning groundwater levels with reasonable accuracy if the reference well is representative for the case study site. The case study is too limited to enable a general conclusion about the accuracy of the method, but the results indicates that there are risks associated with the hydrogeological conceptualization and that there are possible improvements to the description given in TK Geo. Keywords: Groundwater levels, Dimensioning groundwater levels, TK Geo, Hydrogeological conceptualization, Measurement series II Dimensionerande grundvattennivåer En fallstudie av TK Geo-metoden Examensarbete inom masterprogrammet Infrastruktur och miljöteknik IDA NORDBERG Institutionen för arkitektur och samhällsbyggnadsteknik Avdelningen för Geologi och geoteknik Teknisk geologi Chalmers tekniska högskola SAMMANFATTNING Dimensionerande grundvattennivåer är ett vanligt förekommande begrepp inom anläggningsbranschen. Nivån används för att beräkna det tryck som en konstruktion under mark ska motstå. Konceptet med dimensionerande nivåer återfinns även i Eurocode och Trafikverkets styrande dokument och konceptet förutsätter liknande hantering av grundvatten jämfört med andra laster som en konstruktion kan utsättas för. Problemet är dock att grundvattennivåer fluktuerar och att ansätta ett dimensionerande värde är svårt. Men det är inte bara utvärderingen av dimensionerande nivåer som skapar frågetecken i branschen. Även vem som ska vara ansvarig för utredningarna och hur frågan ska hanteras i stora projekt där hela landskapet omvandlas och de hydrogeologiska förutsättningarna förändras totalt är oklart. Det finns inga enkla lösningar till dessa frågor, men i ett försök att bidra till branschen är syftet med denna masteruppsats att genomföra en fallstudie där en utvärderingsmetod för dimensionerande grundvattennivåer testas och jämförs med statistisk analys. Masteruppsatsen inkluderar också en litteraturstudie som beskriver grunderna för grundvattennivåers fluktuerande. Litteraturstudien ger också en överblick kring hur laster och grundvattennivåer behandlas i Eurocode och svenska styrande dokument. Den testade utvärderingsmetoden för dimensionerande grundvattennivåer kommer från styrande dokument från Trafikverket, kallade TK Geo. Metoden är från början utvecklad av Chester Svensson och Göran Sällfors under 1980-talet och är en simpel metod för att ansätta dimensionerande grundvattennivåer för byggplatser. Metoden kallas för TK Geo-metoden i denna rapport. TK Geo- metoden är en referensrörsmetod som ansätter grundvattennivåer på en specifik plats utifrån mätserier från närliggande referensrör. I detta projekt används TK Geo metoden för att uppskatta dimensionerande grundvattennivåer för två observationsplatser. Resultaten jämförs sedan med statistisk analys av mätserier från observationsplatserna. Resultaten visar att TK Geo-metoden kan ansätta dimensionerande grundvattennivåer med tillräcklig precision om referensröret är representativt för observationsplatsen. Fallstudien är dock för begränsad för att några generella slutsatser ska kunna dras, men resultatet visar på risker kopplade till den hydrogeologiska konceptualiseringen och att potential till förbättring av beskrivningen av metoden i TK Geo finns. Nyckelord: Grundvattennivåer, dimensionerande grundvattennivåer, TK Geo, hydrogeologisk konceptualisering, mätserier CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 III Contents ABSTRACT I SAMMANFATTNING II CONTENTS III PREFACE V 1 INTRODUCTION 1 1.1 Background 1 1.2 Aim 2 1.3 Limitations 2 1.4 Research questions 2 2 LITERATURE STUDY 4 2.1 Aquifers 5 2.1.1 Aquifer properties 7 2.2 Groundwater recharge 7 2.2.1 Swedish conditions 9 2.3 Groundwater level variation 9 2.4 Hydrogeological conceptualization 11 2.5 Standards affecting geo-construction 12 2.5.1 Evaluation of characteristic and dimensioning values 13 2.5.2 Comments to Eurocodes approach 14 2.6 Methods for evaluation of groundwater measurements 16 2.6.1 Eurocode 16 2.6.2 TK Geo 17 2.6.3 Svensson & Sällfors 17 2.6.4 Extreme value analysis 19 2.6.5 Calculation method for evaluation of return times 20 3 METHOD 22 3.1 Conceptualization 22 3.2 TK Geo method 24 3.2.1 Selection of reference well 25 3.3 Extreme value analysis 26 4 RESULTS 27 4.1 Lorensberg 27 4.1.1 Land use 27 4.1.2 Geology 28 4.1.3 Hydrogeology 29 CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 IV 4.1.4 Hydrology 30 4.1.5 Observation well and reference well characteristics 31 4.1.6 Dimensioning groundwater levels 33 4.2 Brudaremossen 34 4.2.1 Land use 34 4.2.2 Geology 35 4.2.3 Hydrogeology 37 4.2.4 Hydrology 37 4.2.5 Observation well and reference well characteristics 37 4.2.6 Dimensioning groundwater levels 40 5 DISCUSSION 44 5.1 Case study findings 44 5.2 Uncertainties connected to case study calculations 46 5.3 The required length of the measurement series 47 5.4 Need for hydrogeological conceptualization and selection of suitable reference well 48 5.5 Risks connected to TK Geo method 49 5.6 Further development of TK Geo method 50 6 CONCLUSIONS AND FURTHER STUDIES 52 7 REFERENCES 53 8 APPENDICES 56 8.1 Appendix I 56 8.2 Appendix II 56 CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 V Preface This master’s thesis is the completion of my civil engineering studies at Chalmers Technical University in Göteborg. This master’s thesis would not have been possible without the interest and contribution from several people. Firstly, I want to thank Linn Ödlund Eriksson, my supervisor at Sweco in Göteborg, for the opportunity to do my master thesis at Sweco and for her genuine interest and supervision during the whole thesis project. I am very grateful for our discussions along the way and the time you have spent on this project. I also want to thank Ingvar Rehn at Sweco for his contribution to and opinions about the project. It has been an important part, especially regarding finding a suitable method. Furthermore, I want to thank Johanna Merisalu, my supervisor at Chalmers, for her great support along the process and especially for her help with various report writing issues. Lastly, I want to thank Lars Rosén, my examiner at Chalmers. Finally, I want to thank everyone that have helped me staying motivated during the many weeks of working from home this incredibly special season. Ida Nordberg, Göteborg, June 2020 CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 VI CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 1 1 Introduction This master thesis covers the topic of dimensioning groundwater levels and the assessment of these within the framework of infrastructure projects. The thesis project consists of a literature study and a case study. The literature study explains the basic hydrogeological concepts that affects groundwater levels and the background to assessment of dimensioning levels within infrastructure projects. The case study focuses on testing and evaluation of one Swedish method for assessment of dimensioning groundwater levels. 1.1 Background Hydrogeological issues associated with infrastructure projects has become an increasing part of the everyday work for hydrogeologists. The need for hydrogeological knowledge and understanding in infrastructure projects has increased together with the establishment of environmental legislation, increased awareness of the subject, harder competition between opposing interests, and increased complexity of the projects. This has led to more involvement of hydrogeologists in these projects. As a result, hydrogeologists, consultants, and experts from other disciplines such as geo-construction and geotechnics must collaborate to a greater extent than before since knowledge regarding groundwater level fluctuations has become more important for the design of geo-constructions. Groundwater has of course been managed in infrastructure projects even before hydrogeologists were involved to a greater extend, and it exists methods and instructions for hydrogeological issues in e.g. Eurocode 7 (Svensk Standard, 2009a, 2009b) and governing documents from the Swedish Transport Administration (Trafikverket) (Trafikverket, 2016a, 2016b). The hydrogeological regulations and standards described in these references, however, are written from other perspectives than hydrogeology and it has not been written by hydrogeologists. The lack of hydrogeological perspective in the standards and governing documents may lead to that risks connected to hydrogeology are missed out or underestimated in the planning process for new constructions. Hydrogeological risks could for example be the lack of understanding of how an excavation or underground construction affects groundwater levels, and the consequences of that may be that proper actions are not taken, and thus causes damage to the surroundings (and probably additional costs for the responsible actor). Another example could be the lack of understanding of wells, i.e. what a measured level in a well represents and how measurement series should be treated. The consequences may be incorrect handling of data that leads to inaccurate assessment of governing levels, which may be used for dimensioning of constructions. Consultants working with the design of constructions and geo-constructions need input data in the form of groundwater level assessments to be able to design resilient and safe buildings and foundations. The groundwater level used in these calculations is referred to as dimensioning groundwater level. The concept of dimensioning groundwater levels is defined by Eurocode, but different ways of evaluating the parameter exist in various governing documents and standards. It is not only the evaluation of the parameter that confuses the industry, other questions also have arisen in large infrastructure projects (Forssberg, 2020). One common question is who should assess the dimensioning groundwater level? It is the client or the contractor or someone else? This question comes from the uncertainty regarding if dimensioning CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 2 levels is a precondition or something that is “constructed” to control the project. Together with that follow the question if dimensioning levels that are based on how it looks at the site before the start of the project are relevant at all? The whole surrounding landscape may be converted and the water flow directions and volumes in the ground changes entirely for big infrastructure projects, and a dimensioning level that was assessed regarding the conditions at the site before the project started may be irrelevant. There are many questions and uncertainties connected to the concept of dimensioning groundwater levels and the various problems arise in different projects due to the conditions at the site and the type of project. This thesis project focuses on the evaluation of one common method (the TK Geo method) for assessment of dimensioning groundwater levels and thus try to make a minor contribution to the subject within the industry. 1.2 Aim This master’s thesis aims to evaluate the method for calculating dimensioning groundwater levels presented in TK Geo by performing and comparing case studies. The thesis also aims to examine if there are any hydrogeological risks with the method and give some recommendations to further develop it. 1.3 Limitations The thesis project focuses on dimensioning groundwater levels in soil aquifers. Many closely related questions exist but will not be included in this thesis project. These are: • Definition and calculation methodology for characteristic groundwater levels • Impact of climate change on dimensioning groundwater levels • Climate factors impact on groundwater levels in different parts of the country • Plants or installations effect on groundwater levels or how to calculate dimensioning groundwater levels when the ongoing human activity itself changes the groundwater conditions • Already affected groundwater conditions by e.g. drinking water production or underground constructions • Characteristic and dimensioning groundwater levels in rock. 1.4 Research questions To achieve the aim the following questions must be assessed: Literature study • What is groundwater? What is an aquifer? • How is groundwater recharged? Hydrogeological cycle? Water balance? Recharge and discharge areas? • Which factors affect the natural groundwater level variations? CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 3 • What is hydrogeological characterization? How is hydrogeological characterization performed? • What are characteristic and dimensioning values? How are groundwater levels treated within construction? • How is dimensioning groundwater level treated in Eurocode and TK Geo? • How does the calculation method for dimensioning groundwater levels in TK Geo works? Case Study • How accurately does the method in TK Geo predict the dimensioning groundwater level? • How long measurement series is needed for the observation well and the reference well used in the TK Geo method? • To which extent concerning hydrogeological aspects does the site where the observation well and reference well is located need to be conceptualized? • How can a reference well that suites the observation well be found (needed for the TK Geo method calculations)? • Are there any risks connected by using the method described in TK Geo as it is? • In which ways can the method in TK Geo be developed to increase hydrogeological accuracy and the user-friendliness? CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 4 2 Literature Study In the European Water Framework Directive groundwater is defined as: “Groundwater means all water which is below the surface of the ground in the saturated zone and in direct contact with the ground or subsoil” (Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy [2000] OJ L 327/1). Groundwater is used for many different purposes in society, e.g. public supply of drinking water, irrigation, industrial use, and in the mining industry (Kresic, 2007). Groundwater also has an important part in the hydrologic cycle where it contributes to movements of water between atmosphere, land, and oceans. The hydrologic cycle, see Figure 1, shows how the water moves in space and between phases; ice (solid), liquid and vapours (gas). Figure 1: Hydrological cycle (U.S.G.S, 2005). PD. Groundwater per definition can occur in any soil strata from very fine clay particles to coarse gravel and in bedrock fractures and pores. The different types of soil have very different hydraulic conductivity (Fetter, 2014), from as high as 0.1 m/s for well-sorted gravel to as low as 10-12 m/s for some clays (Carlsson & Gustafsson, 1997). The possibility to extract groundwater from the geologic formations follows in general variations in the hydraulic conductivity (low hydraulic conductivity, low extraction possibilities). A geologic formation that can store and transmit significant quantities of potable groundwater (e.g. to a well) is called an aquifer (Kresic, 2007). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 5 2.1 Aquifers Groundwater can occur in both soil and rock. There are two main types of aquifers; confined and unconfined. A confined aquifer is overlaid by a confining layer, a layer with very low hydraulic conductivity, and thus can the hydraulic head be above the upper boundary of the aquifer (Fetter, 2014). A confined aquifer may be an artesian aquifer if the water head is above the ground surface. An unconfined aquifer, also called a water-table aquifer, has the upper boundary at the water table where the pressure always is at the atmospheric level. An intermediate type of aquifer is leaky aquifers which is an aquifer that is over or underlaid of a layer with low hydraulic conductivity that still can transmit water to or from the aquifer (Carlsson & Gustafsson, 1997). The most common types of aquifers are shown in Figure 2. Groundwater in rock occur in fractures that form a system where water can flow (Kresic, 2007). Aquifers in rock are outside the scope of this thesis and will not be given any further considerations. Figure 2: Different types of aquifers. An aquifer can be divided into several features which also relates to the water flows in the aquifer (Kresic, 2007). The aquifer extent is the surface projection of the boundaries of the aquifer. For an unconfined aquifer, see example in Figure 3, the aquifer extent is simple to derive but for a confined aquifer it can be more difficult. The recharge area is the land surface where the aquifer receives water by percolation of precipitation, surface runoff, and maybe also directly from surface water bodies such as lakes and rivers. In simple cases, the aquifer extent and recharge area are the same, but for more complex situations (e.g. confined aquifers) they may be different. A confined aquifer is usually recharged at the edge where the water-bearing layer reaches the ground while an unconfined aquifer is recharged by percolation over the whole area. The recharge area for a confined aquifer can be a quite small part of the aquifer extent compared to an unconfined aquifer where the two features are represented by the same area. The drainage area is the area from which surface runoff will recharge the aquifer (if the water is not evapotranspirated). The drainage area is sometimes also called the contributing area. The water leaves the aquifer through the discharge area, such as via streams, lakes, wetlands, or oceans or via springs. The different key features of a simple unconfined aquifer can be seen in Figure 3. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 6 Figure 3: Key features of an unconfined aquifer. The key features of the aquifer and the linked water flows are also important for the water balance of the system. A water balance is a summation of the water that enters and discharges a limited area during a given time (SMHI, n.d.-d). A water balance can be set up for a single aquifer or bigger systems. A water balance relates the key features (such as aquifer extent, recharge and discharge area etc.) to different flows of water that contribute to the whole system. The recharge area is the area where the water enters the aquifer and the discharge area where the water leaves the aquifer (Kresic, 2007). The drainage area answers the question where all water comes from and the aquifer extent says something about where it flows. The water balance can also be constructed as an equation, see Equation 1. 𝑅 = 𝑃 − 𝐸𝑇 − ∆𝑆 (1) Where: R=Runoff P=Precipitation ET=Evapotranspiration ΔS=Change in storage Change in storage is the term in the equation that is directly related to the groundwater level since large storage corresponds to a high groundwater level and vice versa. The other terms are also indirectly related to the groundwater levels in an aquifer since it affects the amount of water that is available for recharge. Larger recharge leads to higher levels and the opposite leads to lower levels. According to the equation, the change in storage (e.g. in an aquifer) can be estimated as the precipitation over the drainage area minus the evapotranspiration of water directly on the surface and through plants and the runoff (usually most as surface runoff). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 7 2.1.1 Aquifer properties A common method used for the evaluation of aquifers is test pumping. From test pumping results (drawdown in wells) it is possible to evaluate the transmissivity and storativity of the aquifer (Carlsson & Gustafsson, 1997). Transmissivity and storativity are the properties that are simplest to quantify for an aquifer since they are included in commonly used well-equations such as Theis and Thiems. Transmissivity is defined as the water flow through a section of an aquifer that is one unit wide under a gradient of one (Carlsson & Gustafsson, 1997). The water's density and viscosity must be considered. Transmissivity is usually given in m2/s. Transmissivity is linked to hydraulic conductivity. Hydraulic conductivity is a coefficient that describes at which rate water can move through a porous medium and is usually given in m/s (Fetter, 2014). Hydraulic conductivity is often seen as a material property of the granular material, even if density and viscosity of the water also must be considered. If the hydraulic conductivity is known the transmissivity can be found by multiplying the hydraulic conductivity with the saturated thickness of the aquifer. Storativity is defined as the volume of water an aquifer takes to or releases from storage per unit surface area per unit change in head (Fetter, 2014). These changes depend on the volume change of the water and the grain orientation in a confined aquifer. The volume change in an unconfined aquifer is negligible and the storativity is equal to the specific yield. Specific yield is defined as the volume of water a saturated soil or rock will yield due to gravity drainage given as a ratio to the total volume of the soil or rock (Carlsson & Gustafsson, 1997). Changes in transmissivity and storativity covaries, and a high transmissivity give a high storativity. Anyhow, the magnitudes of the values are different for confined and unconfined aquifers. To enable any conclusions if a specific aquifer has a high or low transmissivity or storativity the aquifer type must be known. 2.2 Groundwater recharge Groundwater recharge is dependent on several factors but many of them are difficult to control or measure. Kresic (2007) lists some factors that affect the groundwater recharge. The first factor is the climate, thus whether the aquifer is located in a humid or arid region influences the groundwater recharge pattern and amount. In humid regions, groundwater recharge is usually dominated by precipitation infiltrating from a large drainage area and with the groundwater levels located close to the surface. Streams and lakes act as drainage areas to the aquifers and are gaining water. In arid regions, aquifer recharge is often concentrated to small areas, such as losing streams and deep water tables are common. Recharge by infiltration is to a great extent controlled by the availability of water at land surface which is limited by climate factors such as precipitation and evapotranspiration (Kresic, 2007). The water balance equation (Equation 1) links the change in storage, which can be the change in groundwater storage for a specific aquifer, to the different climate factors, precipitation, runoff, and evapotranspiration. Precipitation is a crucial factor for groundwater recharge since larger amounts of CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 8 precipitation enable larger amounts of recharge, but not all precipitation recharges the groundwater aquifers. Part of the water is directly trapped by the vegetation and do not reach the ground; this is called interception. Some of the water runs off on the surface, in surface water bodies or groundwater streams close to the surface, and never creates deep groundwater recharge. The water that infiltrates to the unsaturated zone is either percolated further to the saturated zone and thus recharges the groundwater or is taken up by vegetation to facilitate photosynthesis. The latter process is called transpiration. Water can also evaporate directly from the ground. In the water balance equation, evaporation and transpiration is merged together to one factor and is then called evapotranspiration. The evapotranspiration is dependent on the climate (e.g. temperature and soil moisture) and it is important to distinguish between actual and potential evapotranspiration. Actual evapotranspiration is the evapotranspiration that occurs under the given climate conditions and potential evapotranspiration is the one that would occur in the given climate if the soil moisture were unlimited (Fetter, 2014). If the actual evapotranspiration is higher than the precipitation during a period, it leads to limited groundwater recharge and vice versa if the precipitation is higher than the actual evapotranspiration. The second factor that Kresic (2007) lists are land-cover and land use. The infiltration of water into the ground is dependent on the land crusts permeability. Land covered with vegetation has usually higher permeability than bare soil since pores in bare soil more often becomes sealed due to accumulation of fine particles and thus, does not allow for infiltration. The land-use also effects infiltration of water, and urban and developed areas have, in general, lower infiltration rates than rural areas due to the high degree of more or less impermeable surfaces (such as asphalt). The third factor that influences groundwater recharge according to Kresic (2007) is porosity and permeability of soil cover. This factor includes the transmission rate of water through the soil layers above the water table (the unsaturated zone), while the second factor only counts for the soil's crust. The fourth factor is the geologic and geomorphologic characteristics. This factor counts for bedrock characteristics which may have an important role for infiltration if the soil cover is thin or absent. For example, fractured bedrock will allow more infiltration than unfractured crystalline bedrock and steep slopes will increase the surface run-off and thus decrease the infiltration. The fifth and last factor that Kresic (2007) list as a factor that influences groundwater recharge is depth to water table. In arid regions with deep water tables, the infiltration will end long before the wetting front reaches the water table. The water may then be evapotranspirated or stored in the unsaturated zone instead of recharging the groundwater. In general, increased thickness of the unsaturated zone leads to decreased groundwater recharge (Fetter, 2014). The groundwater recharge is also affected by seasonal variations and when the precipitation falls during the year (Fetter, 2014). The groundwater is not recharged when the evapotranspiration is equal to or exceeds the amount of precipitation or when no precipitation falls at all. The evapotranspiration is high during the vegetational season and that also coincides with less precipitation in many temperate areas. In areas where the precipitation falls as snow during the winter, most of the groundwater recharge occurs during the snow melting (SGU, 2017a). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 9 2.2.1 Swedish conditions The ground infiltration capacity is in general high, and therefore the groundwater recharge in Sweden is mostly dependent on other climate factors such as precipitation and evapotranspiration (SGU, 2017a). A rough estimation, that is valid for many locations is that the groundwater recharge can be calculated as the precipitation minus the evapotranspiration. This difference is also called the effective precipitation. Even if precipitation and evapotranspiration are the crucial factors for groundwater recharge in Sweden, the geology and land use also affect the infiltration capacity of the soil. The climate varies throughout Sweden, and this has a large effect on when and to what extent groundwater recharge occurs in the different parts of the country (SGU, 2017a). In large parts of Sweden, the precipitation falls as snow during the winter and the groundwater recharge appears when the snow melts. This creates a peak in groundwater recharge during the spring for the northern parts of Sweden and the level then decreases during the vegetational period (summer). The recharge increases in general in all parts of the country in the autumn due to more precipitation, but it is more significant in the southern parts. During the winter, the groundwater recharge is nearly zero in the northern parts (as the precipitation is stored as snow) but some groundwater recharge occurs in the southern part of the country. The amount of precipitation in Sweden is largest in the southwestern parts of the country and less on the east coast (SGU, 2017a). In the mountains, the precipitation varies a lot locally due to the topography with mountains and valleys. A lot of the precipitation falls during summer and autumn, but it varies between different parts of the country and the differences between the years are also significant. If the precipitation falls during the vegetational period (summer months, the length varies along the country) most of the water evapotranspirates and the recharge becomes practically zero. 2.3 Groundwater level variation Groundwater levels in aquifers vary with seasons during the year but also between years (SGU, 2017a). The level increases during periods when the groundwater is recharged and falls when no precipitation occurs or when the evapotranspiration exceeds the precipitation (Fetter, 2014) as described in the sections above. The groundwater variation is not only dependent on recharge and discharge, it is also depending on the storage capacity in the actual aquifer (SGU, 2017a). The storage capacity is the amount of water that the geological formation can hold. The storage capacity is depending on the effective porosity (the connected pores that allow transmission of water). A geologic formation with high effective porosity has a high storage capacity since the void space that can be filled with water is large, and accordingly a geologic formation with low effective porosity has low storage capacity. The higher the effective porosity is the faster water can be transmitted in the pores (higher transmissivity and also storativity) but it also requires a larger volume of water to fill up the voids, consequently the material has a larger storage capacity (Svensson, 1984). CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 10 An unconfined aquifer with high storage capacity is less sensitive to periods with low or no recharge while unconfined aquifers with low storage capacity can have rapid level change when the recharge decreases or increases (SGU, 2017a). An unconfined aquifer with high storage capacity (large effective porosity) needs a large volume of waters to increase the groundwater level in the whole aquifer, and thus have smaller natural variations than an unconfined aquifer with low storage capacity (Svensson, 1984). In an unconfined aquifer with low storage capacity, the effective porosity rapidly gets filled with water after a precipitation event and the level rises, while during a dry period, the groundwater levels quickly gets lower. An unconfined aquifer with low storage capacity accordingly has larger fluctuations in groundwater levels during a year. A confined aquifer is characterised by the pressure level (that may be above the upper boundary of the water bearing material) and the volume of water that is needed to change the pressure level is not the same amount of water that is needed to rise the water table in an unconfined aquifer. Within a watershed, higher locations in the terrain (recharge areas) in general shows larger groundwater variations than lower located areas (Svensson, 1984). Higher located areas often show a fast reaction to precipitation and dry periods, and sometimes wells in such locations become dry during long periods without precipitation. The groundwater level is not only a subject for natural variation, it can also be affected by human activities such as extraction for drinking water production or for industrial activities or change due to leaking underground constructions. The effects caused by human activities may be much greater than natural variations in some aquifers. The Geological Survey of Sweden (Sveriges geologiska undersökning, SGU) is the responsible authority for groundwater related issues in Sweden including measurements and assessments of levels. When SGU presents their assessments of the groundwater situation in Sweden they distinguish between small and large aquifers (SGU, n.d.-b). SGU defines small aquifers as all aquifers that has an area of 5 km2 or less and accordingly, large aquifers are all aquifers that has an area that is larger than 5 km2 (SGU, 2017b). Small aquifers react usually fast to changes in precipitation due to their limited storage capacity (SGU, 2017a). Changes can often be noticed as soon as just a couple of days after precipitation. In addition, aquifers with low storage capacity react rapidly to changes in precipitation even if the aquifers may have a big areal spreading. Such aquifers can be found in till and crystalline bedrock (materials with low effective porosity and thus a quick response to infiltration of surface water). Large aquifers react slowly to changes in precipitation. It can take weeks or months before a change can be observed (due to the large volume of water that is needed to fill the voids in the aquifer and thus raise the water level) (SGU, 2017a). Large aquifers are found in large sand and gravel formations such as eskers and deltas. A large aquifer is not as sensitive for shorter periods with less groundwater recharge as small aquifers, but if the recharge is limited over several years it may have a significant effect. If the groundwater levels decrease in a large aquifer it takes long time before it reaches its normal levels again. Large aquifers are important raw water sources for municipal drinking water production. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 11 2.4 Hydrogeological conceptualization Hydrogeological conceptualization of an aquifer is a way to describe the properties and interactions between surface- and groundwater bodies (Fulton et al., 2005). One way to describe the characteristics of an aquifer is to create a conceptual model of the system. The conceptual model should include all necessary information that is needed to understand the system, its connections to extern events and the problem considered, but not include unnecessary information. A conceptual model may contain information about climate parameters (e.g. precipitation, evapotranspiration), geology, topography, land use, hydrogeologic characteristics (e.g. groundwater recharge, flow directions) and water-quality parameters, but it depends on the purpose with the model and information available. Which information that is relevant to include in a conceptual model depends on the local conditions at the site and the purpose with the model. Stejmar Eklund (2002) describes which factors to include in a model of a site with focus on Swedish conditions. According to Stejmar Eklund, a hydrogeologic environment can be described by two systems, a hydrogeologic system, and a parameter system. The hydrogeologic system should include information about the bedrock, soil, location of the site relative to the highest shoreline, stratigraphy, hydrology, and land use. If the site is located above or below the highest shoreline gives information about possible stratigraphic units underlaying the surface crust. Geological formations formed below the highest shoreline can include both glacial and postglacial clay layers and other postglacial sediments such as wave-washed sands and peat above the till and glaciofluvial sediments. Stratigraphic formed above the highest shoreline includes no clay layers above the layers with till and glaciofluvial sediments, but may be overlaid with peat or different types of sorted sediments such as wind sediments, see Figure 4. These two main stratigraphic are valid for most of Sweden, but the thickness of the layers can vary a lot, and all layers may not be represented at all sites. Figure 4: Main stratigraphy’s in Sweden, above highest shoreline to the left and below highest shoreline to the right. The parameter system should include data about groundwater recharge, groundwater level, hydraulic conductivity, porosity, storage capacity, field capacity, and the hydraulic gradient (Stejmar Eklund, 2002). CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 12 In some situations, a characterization of the environment around a specific well may be of interest. When characterizing a specific well, all factors that describe the whole system mentioned in the paragraphs above are of interest. Beside these facts, information about where in the aquifer the specific well and well screen is situated also are of interest since the water level in the well represents the pressure level at the well-screen location. Especially, information regarding which aquifer type (confined or unconfined) the well is situated in and if the well is in a recharge or discharge area is important. 2.5 Standards affecting geo-construction It is crucial to have a proper foundation when building infrastructure such as roads, railways, bridges, tunnels, or pipes. The foundations should resist both the loads from the structure itself and loads associated with activities. The strength of a foundations is always dependent on the soil underneath and the geotechnical conditions at the site. The design of infrastructure constructions is affected by various standards and rules e.g. Eurocode. Eurocode is a European wide regulation and consists of 10 standards that cover topics related to structural design and civil engineering (Svensk Standard, 2010). The first part, only called Eurocode (SS-EN 1990), sets the basics for structural design and defines terms used in all parts. Geotechnical design is more in-depth examined in Eurocode 7 (part 1 and 2). In addition to Eurocode, The Swedish Transport Administration (Trafikverket) have their governing documents that are mandatory in every project purchased by them (Trafikverket, 2016a). For geotechnical constructions, the governing document is TK Geo (Trafikverkets tekniska krav för geokonstruktioner). In the first chapter of TK Geo, the Swedish Transport Administration state that dimensioning of geotechnical structures should be done according to Eurocode 7, but their document adds some complementary rules to follow for Swedish conditions. Together with TK Geo, the Swedish Transport Administration has issued TR Geo (Trafikverkets tekniska råd för geokonstruktioner) that adds advice to the requirements stated in TK Geo (Trafikverket, 2016b). Eurocode requires that a structure should be designed and performed in such a way that it resists all likely actions (loads) during its intended lifetime with satisfying reliability and that it remains fit for its intended use (Svensk Standard, 2010). The design of structures according to Eurocode is based on the limit state principle, and Eurocode defines two different limit states. Firstly, the Ultimate Limit State (ULS), that is defined as a limit state that concerns the safety of people and the structure. Secondly, Serviceability Limit State (SLS), which is a limit state that concerns the functionality of the structure, comfort of people, and appearance of the construction work. Eurocode requires that the design is verified so that any relevant limit state not is exceeded during the lifetime of the structure and in any specified design situation. Eurocode defines design situations as conditions in which the structure can be during different moments of its lifetime (Bond & Harris, 2008). In Eurocode, four different design situations are described (Svensk Standard, 2010). The four design situations are; persistent design situations (where the structure is during normal use), transient design situations (temporary conditions), accidental design situations, and seismic design situations. All design situations are connected to different actions (load situations) and every design situation that is relevant to the circumstances under which the structure must fulfil its function should be included in the verification. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 13 In Eurocode, actions are defined as all forces that act on the structure, in other words, all loads the structure needs to resist (Svensk Standard, 2010). Further, a geotechnical action in Eurocode is defined as an action to the structure caused by soil, fill, or groundwater. How to evaluate a geotechnical load is determined by Eurocode 7. In Eurocode 7, several examples of what to include as actions are given and one of them is groundwater pressure (Svensk Standard, 2009a). However, later in the same chapter it is stated that groundwater levels are treated as geometrical data. So, it is a differentiation in the standards between the handling of groundwater pressure as a force (load) to the structure and groundwater levels used as a nominal input parameter. However, the two parameters are dependent on each other since the magnitude of the groundwater pressure to a structure is determined by the location of the groundwater level. Groundwater pressure is also treated as a load in TK Geo (Trafikverket, 2016a). 2.5.1 Evaluation of characteristic and dimensioning values Suitable values of geotechnical parameters are needed for design of geo-constructions (Bond & Harris, 2008). The process of converting test results to design values is called ground characterization and is used for geotechnical parameters in Eurocode. The process is graphically shown in Figure 5. Figure 5: Ground Characterization Process. Firstly, test results obtained from field tests or laboratory tests are used to derive different geotechnical parameters (Bond & Harris, 2008). The derivation can be done either through correlations, theories, or empirical rules. The second step in the process is to evaluate a characteristic value for the specific parameter. A characteristic value of an action is in Eurocode defined as “principal representative value of an action” (Svensk Standard, 2010). The evaluation of a characteristic value should be done with statistical methods as either the 5th and 95th fractile value depending on if a high or low value is most unfavourable. This approach works well for man-made materials such as concrete and steel that is well controlled and statistically known, but falls CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 14 short for geomaterials due to their huge variability in the properties (Bond & Harris, 2008). Therefore, an additional definition of characteristic values for geomaterials are given in Eurocode 7 as “The characteristic value of a geotechnical parameter shall be selected as a cautious estimate of the value affecting the occurrence of the limit state.” (Svensk Standard, 2009a). This principle does not say how to evaluate any specific geotechnical parameter but Bond & Harris (2008) give their interpretation of what a cautious estimate is in their book. Bond & Harris states that a cautious estimate is an “approximate calculation/judgment that is careful to avoid problems or dangers”. Further, Bond & Harris explains the impact of Eurocodes quotation “affecting … the limit state” which means that the characteristic value of a geotechnical parameter is depending on the limit state, and thus exists not any value that are valid for all situations. The characteristic value of a geotechnical parameter varies depending on which limit state that is evaluated. Eurocode also defines characteristic values for geometrical data, such as groundwater levels. Eurocode states that characteristic values of geometrical data should be measured, nominal, or estimated upper or lower levels (Svensk Standard, 2009a). The last step in the ground characterization process is to obtain the design value. The design value is obtained by factorization of the characteristic value. Factorization is done by applying a partial factor to the characteristic value to decrease the risks in the design (Bond & Harris, 2008). Eurocode give special considerations for design values of groundwater pressure. Eurocode states that design values for ultimate limit states shall represent the most unfavourable condition that could occur during the intended lifetime of the structure (Svensk Standard, 2009a). Furthermore, the design values for serviceability limit states should represent the most unfavourable condition that could occur during normal circumstances. Eurocode give two alternative ways to derive design values for groundwater pressure, either by apply partial factors to characteristic water pressure or adding a safety margin to the characteristic groundwater levels. The approach with applying partial factors to groundwater pressures is debated by engineers (Bond & Harris, 2008). An argument not to apply partial factors to water pressures (or water levels) is that the quantity is relatively well known, and it is illogical to apply a safety factor to a known parameter, especially if the highest possible water level is used and it is placed at the ground surface. On the other hand, it is illogical to treat groundwater differently to other loads and it may cause problems in numerical analysis if earth pressures are factored but not groundwater. Eurocode does not give any regulations or recommendations regarding which approach to use when deriving design values for groundwater pressure. Accordingly, it is a question all engineers need to face for every single case. The choice of approach for evaluation of design values may also be dependent on how the selection of characteristic water levels are done. 2.5.2 Comments to Eurocodes approach The partial factor approach is a method that is used for estimation of design values for all types of geotechnical parameters in Eurocode since it simply adds a safety margin (in general partial factors scale up the loads and scale down the resistance/strength CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 15 and if the resistance/strength still is larger than the loads should the construction be safe). This works well for most geo-parameters evaluated with field or laboratory test such as strength of soil or rock and for loads such as traffic loads. However, when it comes to groundwater it starts to be somewhat confusing and unclear how the evaluation should be performed. That depends on several aspects. Firstly, Eurocode write both about managing of groundwater pressures and groundwater levels and states that groundwater pressures are an example of a geotechnical load to a construction while groundwater levels are a nominal geometrical input data. This is true, but the two quantities are also depending on each other. The magnitude of the groundwater pressure to a construction is decided by the groundwater level in the surroundings. Accordingly, to enable calculation of characteristic values of groundwater pressures (that later can be factorized to design values) the groundwater level must be known. Groundwater levels, as describe more in detail in section 2.3, are a result of natural variation and does not take one single value at all time. Groundwater levels varies according to climate (precipitation etc.), with seasons and between years. Consequently, the groundwater pressures to constructions also varies. However, when designing constructions, a decided value for these quantities are needed to enable verification if the construction resists the loads or not. To be able to find the characteristic groundwater pressure, the characteristic groundwater level has to be evaluated first. Characteristic groundwater levels are evaluated from measurement series (more about that in the following section), since one single measurement in an observation well simply tells the status that single day and says nothing about if the level is high or low. From such groundwater measurement series, levels that corresponds to both ULS and SLS has to be evaluated (since structures, in general, must be verified for failure according to both limit states). The evaluated levels can then be used to calculate corresponding water pressures that are needed for estimation of design values (that can be done with either partial factors or by adding a safety margin). Eurocodes procedure with first evaluate characteristic value and then recalculate it to a design value are the most common way to evaluate design values for geotechnical parameter, but dimensioning groundwater levels (i.e. design values for groundwater levels) can also be evaluated directly by using statistical methods (without first asses a characteristic value). In such cases, statistical methods that includes the concept of return times are used, and that are described more in section 2.6 and corresponding sub-sections. The connections between groundwater levels, groundwater pressures and evaluation of corresponding characteristic and design values are not completely understandable when reading in Eurocode. The confusion depends partly on the fact that the information are given in different standards and chapters, but also on the fact that Eurocode not explicitly explain the natural connection between groundwater pressures and groundwater levels and that makes it harder to understand the standards and how it should be used. Figure 6 shows an overview of governing regulations, applicable methods for calculation of dimensioning groundwater levels and the reference/sources to the different methods. Extreme value analysis and a reference well method are described in section 2.6 and corresponding sub-sections. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 16 Figure 6: Overview of governing regulations, methods and references/sources. 2.6 Methods for evaluation of groundwater measurements In the following sub-sections different approaches for evaluating groundwater measurements are described. The methods are used to estimate dimensioning values of groundwater levels needed as input to later calculations in the design process. 2.6.1 Eurocode Eurocode states that evaluation of groundwater level measurements must consider geological and geotechnical conditions at the site, the accuracy in the measurements, changes due to time, the length of the measurement period, the season when the measurements are done and other climate factors during and before the measurement period (Svensk Standard, 2009b). Further on, the evaluated results should include the highest and lowest levels measured during the measurement period and, if possible, should both upper and lower limits for both normal and extreme conditions be derived from the measurement series (correlates to ULS and SLS as described in the comments above). Such evaluation is usually difficult to do since it requires long measurement series which is uncommon to have for infrastructure projects. Eurocode therefore suggests in an informative appendix (compare to normative paragraphs or appendices) that a statistical method can be used to evaluate groundwater measurements instead. A statistical method can be used when only a short measurement series for the construction site is available, but a reference well with a long measurement series in the region exists (Svensk Standard, 2009b). If a short measurement series at the actual site (at least 3 months with at least 7 recorded levels) are available and a 15 years long measurement series (at least) in a reference well that is situated in a similar aquifer exists it is possible to evaluate the dimensioning water levels at the site with statistical methods. Statistical evaluation should be compared with hydrogeological information about the site and the reference site. Eurocode does not explicitly explain the methodology for such statistical methods but refers, in the informative appendix mentioned above, further to Chester Svensson’s publication from 1984 regarding the analysis of groundwater observations (Svensson, 1984). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 17 2.6.2 TK Geo The Swedish Transport Administration states in TK Geo that dimensioning water pressure (a dimensioning value in the Swedish governing documents is the same as a design value in Eurocode) should be evaluated from the most unfavourable water level with at least 50 years return time, but if the consequences are considerable should longer return times be used (Trafikverket, 2016a). Further on the Swedish Transport Administration state that extreme value analysis on the measurement series from the construction site should be used primarily, but if the amount of measurements are limited, they suggest that a method including a reference well could be used instead (the same method as Svensson & Sällfors published already 1985, see further in chapter 2.6.3) or that the value of the dimensioning water levels can be evaluated based on the topography and hydrogeology of the area. TR Geo does not give any requirements on how long the measurement series for the well at the construction site (observation well) or the reference well needs to be if reference well method is used for evaluation of dimensioning groundwater levels. However, they state that the reference well should be located within 50 km from the construction site and in the same climate zone (Trafikverket, 2016b). If these requirements are fulfilled, Equation 2 can be used to calculate the dimensioning level with 50 years return time at the construction site, 𝑌𝑂,𝑚𝑎𝑥,50. 𝑌𝑂,𝑚𝑎𝑥,50 = 𝑌𝑂,𝑚𝑎𝑥 + 𝑆𝑅 50 ∗ 𝑟0 𝑟𝑅 (2) Where 𝑌𝑂,𝑚𝑎𝑥 = maximum level in the observation well during the observation time 𝑆𝑅 50 = the difference between 𝑌𝑅,𝑚𝑎𝑥,50 (calculated maximum value with 50 years return time for the reference well, see chapter 2.6.5 for calculation procedure) and 𝑌𝑅,𝑚𝑎𝑥 (measured maximum level in the reference well) during the observation time 𝑟0 = range (maximum-minimum value) in the observation well during the observation time 𝑟𝑅 = range (maximum-minimum value) in the reference well during the observation time The same equation can be used to calculate dimensioning water levels with longer return times if a lower risk for exceedance is required, by changing the value of 𝑌𝑅,𝑚𝑎𝑥 to another theoretical calculated return time for the reference well. The minimum levels can also be calculated with the same equation if the minimum level with required return time is inserted instead. This, however, is not described explicitly in TK Geo. If Equation 2 is used the dimensioning groundwater levels (design values) are evaluated directly from the measurement series without having to first evaluate characteristic values. The safety is instead counted for when selecting return time used in the calculations. 2.6.3 Svensson & Sällfors The equation for estimation of dimensioning groundwater levels presented in TK Geo (Equation 2) is the same as Svensson & Sällfors published 1985 in a report regarding calculations of dimensioning groundwater levels for construction sites in the Gothenburg area. The equation is similar to an equation Svensson presents in his CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 18 doctoral thesis in 1984. The difference is that Svensson only has done calculations for the existing length of measurement series from reference wells (i.e. he has not estimated levels with longer return time than the existing measurement series length, such as levels with 50 or 100 years return time), which by this time was around 13 years. Svensson & Sällfors developed the method to enable estimation of dimensioning groundwater levels for construction sites in a simple way when only a short measurement series is available for the construction site. In the publications (Svensson, 1984; Svensson & Sällfors, 1985) more explanations are given to the factors in the equation and assumptions and theoretical background to the whole method compared to in TK Geo and TR Geo (Trafikverket, 2016a, 2016b). Svensson & Sällfors also gives a stepwise calculation procedure and show how to use the equation with an example. The method includes some assumptions that have been verified by Svensson (1984) in his doctoral thesis. The first assumption is that the groundwater levels in aquifers, located close to each other, covary (i.e. aquifers in the same region with similar climate such as precipitation etc.). Even if the magnitude (how much the level increase or decrease due to a specific weather event) of the fluctuations varies between different aquifers, the variation tends to be analogous in time (increase and decrease simultaneously). The second assumption is that variations in groundwater levels during a short period are proportional against the total variation width for a specific aquifer. These two assumptions have been verified by Svensson (1984) by systematic test the method with several wells from three different areas in western Sweden and some longer measurement series from other places in Sweden and abroad. The result from the verification of the method shows that it can be used with desirable accuracy (±0.5 m should be reasonable accuracy for geo-constructions (Svensson, 1984)) for calculation of dimensioning groundwater levels for construction projects. The verification also shows that the method predicts dimensioning maximum levels better than minimum levels. That is probably due to the short duration of dry periods and thus a higher risk that these are not registered. Svensson & Sällfors (1985) recommends that the groundwater level in the observation well is measured two times per month for at least three months (seven observations). The selected reference well should be located in the same climate zone as the observation well and for the Gothenburg area is that approximately within a distance of 50 km. The accuracy in the prediction will increase if the reference well and observation well have similar locations within the aquifers, such as high or low position (recharge/discharge area) (Svensson & Sällfors, 1985). This method, where Equation 2 is used to predict the dimensioning groundwater levels, that are developed by Svensson & Sällfors, and implemented in the governing documents by the Swedish Transport Administration, and used and tested in this thesis project is further on in this report noted the TK Geo method, for simplicity. Figure 7 shows an overview of key aspects and terms included in the TK Geo method and how the different wells are related to each other. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 19 Figure 7: Overview of key aspects of the TK Geo method. To use the equation for dimensioning groundwater levels (Equation 2) data from the selected reference well must be evaluated statistically to find the level with a certain return time. This is further explained in the coming sections. 2.6.4 Extreme value analysis Extreme value analysis is from the beginning used to analyse extreme values and the probability of their occurrence (Gumbel, 1958). The probability of occurrence of events with severe consequences for society is important knowledge in many fields and has been for a long time. Especially for businesses that are dependent on constant (or periodical) water supply, such as agriculture, hydroelectric power production, and all types of industries that need large volumes of water for production or cooling. Water supply from surface water bodies is depending on the levels/flow in the natural body and that varies with dry and wet periods. A long dry period may decrease the flow in a river or water level in a lake significantly and thus also the amount of water that can be pumped for societal use. For a business owner it may be important to know how often the water level or flow is too low to allow pumping of water to be able to plan for reserve water or set up emergency plans. For society it is important to know the frequency of high flow or levels to plan for flooding, but also the frequency of low levels or flow in raw water bodies for drinking water production. Extreme value analysis has been used in hydrology for a long time to estimate the probability that a specific event occurs (Gumbel, 1958). The method is commonly used to calculate flows or levels (high and low) in streams and rivers with certain return times but can also be used for precipitation to calculate the intensity e.g. a 100 years rain. Return time is a concept that describes how common a specific event is and it is used to describe how often extremes occur in nature (such as floods or droughts) (SMHI, n.d.-b). The return period is the average time between similar events (Wilson, 1990) e.g. the return time of a certain high level in a lake is 10 years. It means that this specific level will be equalled or exceeded on average once every 10 years. The probability that the same level will be equalled or exceeded this year is 10 CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 20 % and the probability that the level is equalled or exceeds during a specific 10-years period is 65 % since the risk accumulates. Return time and the probability of exceedance is linked by a simple mathematic connection, see Equation 3. 𝑇𝑟 = 1 𝑃(𝑋≥𝑥) (3) Where 𝑇𝑟= Return time 𝑃(𝑋 ≥ 𝑥)= Probability of exceedance Extreme value analysis can be done in two different ways, with series of annual maximum (or minimum) values or with exceedance series (Svensson, 1984). For annual maximum series, the maximum value each year (or hydrogeologic year) is selected. These values form a maximum series that can be adjusted to common statistical distributions and levels with longer return time than the length of the series can be calculated. In exceedance series, all values above a selected threshold value is included. This usually gives series with more data to adjust to a statistical distribution. However, groundwater levels are not independent which is one of the fundamental requirements in statistical analysis, therefore, the use of exceedance series for groundwater data is less appropriate. One way to come around this problem is to use some declustering method to ensure the independence of the values included in the series (Haaf, 2015). The next question then becomes which distribution the extreme value series should be adjusted to? Svensson (1984) tested to adjust the extreme value series of groundwater levels to five common distributions and found that normal distribution cannot be rejected for groundwater levels. The normal distribution is also used in the calculation method for dimensioning groundwater levels by Sällfors & Svensson (1985) to find the level with a specified return time for the reference well. Haaf (2015) has tested the adjustment of groundwater level series to more distributions and found that it depends on the measurement series (i.e. the form of the maximum/minimum values distribution, such as tails etc.) which distribution that fits best. Haaf further finds that normal distribution usually cannot be rejected for most of the series (same as Svensson found), but also that generalized extreme value distribution and log Pearson type III distribution have a good fit for many series. 2.6.5 Calculation method for evaluation of return times In the publication by Svensson & Sällfors (1985), a stepwise procedure to statistically treat the data from the reference well is described. The procedure is used to find the level with a certain return time needed in the term 𝑆𝑅 in Equation 2. To enable prognosis of high and low levels with long return times (usually much longer than the length of available measurement series) for the reference well is the normal distribution used (Svensson & Sällfors, 1985). Firstly, the maximum and minimum values of each year are selected, and these values form the series of annual maximum and minimum values. The method uses hydrologic years as base. The hydrologic year in southern Sweden is from 1 October to 30 September. The values in the annual maximum and minimum series are then ranked from most extreme to less extreme (highest to lowest for maximum values and reverse for minimum values) by giving the values a rank number starting from 1. The rank number is denoted m. For CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 21 each extreme value, the probability of exceedance is calculated with Weibull’s formula, see Equation 4. 𝑃 = 𝑁+1−𝑚 𝑁+1 (4) Where N=total number of values in the extreme value series m=rank of the value The probability is then plotted against the maximum (or minimum) value on a normal probability paper, see Figure 8 for an example of how it looks in theory. From this graph, the levels corresponding to longer return times could be extracted (since a specific return time corresponds to a specific probability of exceedance). Figure 8: Theoretical plot of groundwater levels on probability paper. The level with a certain return time can also be calculated based on the annual extreme value series with Equation 5 (Svensson & Sällfors, 1985). 𝑦𝑚𝑎𝑥/𝑚𝑖𝑛 𝑇 = �̅�𝑚𝑎𝑥/𝑚𝑖𝑛 ± 𝑡𝑇𝑠𝑚𝑎𝑥/𝑚𝑖𝑛 (5) Where 𝑦𝑚𝑎𝑥/𝑚𝑖𝑛 𝑇 = The value of a level with the certain return time, either max or min �̅�𝑚𝑎𝑥/𝑚𝑖𝑛 = average of the values in the annual extreme value series 𝑡𝑇 = frequency factor (see Appendix I) 𝑠𝑚𝑎𝑥/𝑚𝑖𝑛 = standard deviation of the values in the annual extreme value series The calculated level with a certain return time is then used in factor 𝑆𝑅 to calculate the dimensioning groundwater level. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 22 3 Method The method is developed to answer the research questions stated for this thesis project and to fulfil the aim. The aim is to test the accuracy of the calculation method for dimensioning groundwater levels presented in TK Geo by performing a case study and investigate if there are any hydrogeological risks connected to the calculation method. Within the case study was the method tested on two different locations, one in an urban environment and one in a rural environment. The locations for the case study were mainly chosen by the ability to find sites where long measurement series of groundwater levels are available. The method includes three different parts (see Figure 9); conceptualization of the sites, calculation of dimensioning groundwater levels according to TK Geo, and calculation of dimensioning groundwater levels with the statistical method extreme value analysis. The TK Geo method calculations were performed with three different observation period lengths to analyse if the length affects the accuracy of the results. The different parts are described more in detail in the sections below. Figure 9: Overview of the method. The results from the two calculations parts were then compared to evaluate how precise the TK Geo method estimates the dimensioning groundwater levels. 3.1 Conceptualization The first step was to conceptualize the selected site to understand the hydrogeologic conditions at the site and the position of the well that the measurements come from. The conceptualization was done by gather available information about the site from different sources, see Figure 10 for sources used in different parts. How detailed the conceptualization can be done depends on the information available for the specific site and that varies for the sites used in the case study. In section 2.4 in the literature study, hydrogeological characterization is described in general and factors that are important to include in Swedish settings according to Stejmar Eklund (2002) are listed. Information about all properties in the hydrogeological and parameter systems may not be possible to find for all sites included in the case study. The focus in the conceptualization of the sites were on general information about geology, hydrogeology, hydrology, and land use. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 23 Figure 10: Information sources used for the conceptualization of sites. The conceptualization was performed as a desk study where information about geology, hydrogeology, hydrology, and land use was gathered. The first step was to look at maps, both topographic and geological maps. Topographic maps were used to get an understanding of the location of the site compared to the surroundings and get knowledge about land use. Geological maps were used to get information about soil and water in the ground. The different types of geological maps used for characterization of the geology at the sites were quaternary deposits, soil depth, and rock type. Maps that show the spreading and extraction possibilities of groundwater in aquifers were used to characterize the hydrogeology of the site. Used maps come from The Swedish National Survey (Lantmäteriet) for the topographic map and the Swedish Geological Survey (Sveriges Geologiska Undersökning, SGU) for geological maps. The well archive (Brunnsarkivet) was used to search for drilled wells in the surroundings. The well archive provided by SGU is an archive for all drilled wells in Sweden, both wells drilled for drinking water supply and wells for energy production (SGU, n.d.-a). Information from drillings in the surroundings to the case study site was used in the characterization of the geology of the sites. Lastly, a map that shows surface water basins were used. The map is provided by The Swedish Meteorological and Hydrological Institute (Sveriges meterologiska och hyderologiska institut, SMHI). From such a map it is possible to find surface water dividers and see the size of the water basin that the case study site is located in. Surface water dividers follow the topography in the landscape and also groundwater levels in fine soils tend to follow the topography while groundwater levels in coarser material (material with higher hydraulic conductivity) tend to follow the bedrock topography instead (Bovin, 2011). The second part of the conceptualization was to look at reported results and findings from historical site investigation and other types of old reports regarding the site. This was done to increase the level of details in the conceptualization and get more site- CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 24 specific information. Sources used for this part varies due to performed investigation at the site and availability of reports. Examples of site investigation that can give valuable information are drillings (e.g. for wells) and geotechnical soundings (e.g. CPT sounding). Protocols from this type of investigation can give information about soil types in different layers, the thickness of layers, depth to bedrock, and hydraulic properties of the layers. Important hydrogeological parameters are hydraulic conductivity and storativity. These were determined by using data from historical site investigations or if no such data was available were numbers from the literature for the soil types present at the site used instead. 3.2 TK Geo method The second part of the case study was to estimate the dimensioning groundwater level with the TK Geo method. The calculations were done with the equation presented in TK Geo (Equation 2, see literature study section 2.6.2). The processing of data was done according to the description of the method by Svensson & Sällfors, (1985), see literature study section 2.6.3., 2.6.4, and 2.6.5. Three different lengths of observation periods were used in the calculations to fulfil the aim of this thesis and test the ability of the method presented in TK Geo to predict the dimensioning groundwater levels. The tested observation periods were 3, 6, and 12 months. In real construction projects, the observation period is the time when the observation well has been measured, but in this thesis project can the period seen as the observation period be chosen by other criteria. The frequency of the performed measurements in the wells located at the sites used in the case study has varied a lot during the years. Some years it is only 2-3 recorded levels and some years have automatic pressure loggers been installed that register the level several times per day (depending on the installation of the logger). For the reference and observation well, the observation period is used to find the range of the groundwater levels (maximum-minimum value during observation period). For the reference well, the observation period also is used to find how the levels during this short period relate to the level with a certain return time (Svensson & Sällfors, 1985). To enable that these estimations become as good as possible, the measurements from both wells should in the very best case be performed on the same day. This is not always obtainable in reality and a difference of a few days may be ok and will not affect the results significantly. If the number of days between the measurement occasion is longer can linear interpolation between available measurement points be used to estimate the level a certain day. This thesis project, where the whole measurement series are available from the beginning, has the advantage to choose which period to use as observation period and has thus select periods with enough resolution of data. If some single level measurement was missed during a used period was linear interpolation used. The method requires the use of a reference well located in the same climate zone. Svensson & Sällfors (1985) states that in the Gothenburg area should that correspond to a distance no longer than 50 km between the observation well and the reference well. Svensson & Sällfors states further that the reliability of the prognosis increases if the wells have similar locations within their aquifers. Important features to know about the well location are if the well is situated in a confined or unconfined aquifer CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 25 and if the well is placed in a recharge or discharge area of the aquifer. The placing of the observation well was interpreted by looking at maps; both topographic and geological to find where the bedrock outcrops, where the soil is permeable and where not, and how the landscape look (heights and valleys). Wells from SGU was used as reference wells and SGU usually provide some information about the well, such as type of aquifer (confined/unconfined), recharge or discharge area, and the depth of the well. Such information is useful in the selection of a representative reference well. 3.2.1 Selection of reference well SGU has three different areas with recorded groundwater wells around Gothenburg, see Figure 11. The wells in these areas have been measured since the early seventies (SGU, n.d.-f). The used reference wells were selected from the wells in these three areas. Figure 11: SGU:s Groundwater wells in the Gothenburg area (SGU, n.d.-f). © Sveriges Geologiska Undersökning. The area north of Gothenburg located south-west of Kungälv has area number 53 (SGU, n.d.-f). The area constitutes only of wells in confined aquifers. The area east of Gothenburg located east of Lerum has area number 54. The area includes three well in unconfined aquifers and five in confined aquifers. The area south of Gothenburg located north of Kungsbacka in Sandsjöbacka nature reserve has area number 52. The CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 26 area includes one well in an unconfined aquifer and the rest are placed in confined aquafers. All three areas are rural. SGU measures in general manually two times per month in all these wells, but in some wells are automatically pressure loggers installed that measure several times per day. The measurement interval may have varied in different wells during the years. To find a reference well that is unaffected by unforeseen events and to the greatest possible extent only varies naturally was all the measurements series from all SGU wells in Kungälv, Lerum and Kungsbacka plotted and inspected visually to search for inaccuracy. The visual inspection was done primarily to find out if the wells were affected by human activities in the surroundings. If some well shows clear signs of being affected or any other strange behaviour was that well deselected for the case study. The next step was to find a well that is representative for the specific case study sites. This was done by comparing the characteristics of the SGU wells to the observation wells and a similar was chosen. The comparison was carried out by listing the main characteristics in a table and then was a representative one selected. The well characteristics included in the comparison are listed in Table 1. Table 1: Well characteristics included in the selection of reference well. Well characteristics Aquifer type (confined or unconfined) Well placing in the aquifer (recharge/discharge) Soil type at the surface (Geological map showing Quaternary deposits) Soil type at the filter To further analyse the representativeness of the selected reference well was the measured levels during the observation period from both the reference and observation wells plotted and the trends were analysed. If both wells covary (increase or decrease in level roughly at the same time etc) was the chosen reference well considered representative for the observation site. If not, another well was selected instead. This imply that optimal (or as optimal as possible) conditions for the case study. In real cases it may be more difficult to find a suitable reference well, and less covariance must be accepted to enable any calculations at all. The analysis in this thesis project was done based on calendar years and not the hydrological year. The effects of that are examined further in the discussion sections in this report. 3.3 Extreme value analysis Extreme value analysis was carried out and the results were compared to the results obtained with the TK Geo method. The whole measurement series from the observation well was used for the extreme value analysis. The calculations were done in the same way as the statistical treatment of the reference well. See the literature study section 2.6.5. The extreme value analysis based on long measurement series is considered to give dimensioning levels as close to the “true values” as possible. If the TK Geo method give dimensioning levels close to the results obtained with the extreme value analysis it is considered as a good result/fit. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 27 4 Results In this chapter, the results from the case study are presented and discussed briefly, a more detailed discussion about risks and further development of the method follows in the next chapter. The results for each case study area are presented in a separate sub-chapter (4.1-4.2). Firstly, the site is described conceptually and then the results of the calculations of dimensioning groundwater level with the TK Geo method and extreme value analysis are presented. 4.1 Lorensberg The first case study site is Lorensberg. In the following sections, the conceptualization of the site (section 4.1.1 to 4.1.4) and the results of the calculations (section 4.1.5 to 4.1.6) are presented. 4.1.1 Land use The case study site Lorensberg is located in the central parts of Gothenburg. The well used as observation well is placed in a park area called Himlabacken, west of Götaplatsen, see Figure 12. The park is surrounded by streets, parking lots, and buildings, thus a lot of impermeable surfaces. Figure 12: Topographic map of the central parts of Gothenburg with the case study location marked with a red circle. © Lantmäteriet. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 28 The well used as observation well for this site is a wells from the city’s groundwater control program. The well is notated GW485 and has been measured since June 1974. The measurement interval has varied during this period. 4.1.2 Geology The soil close to the ground surface around the observation well is postglacial silt (SGU, n.d.-g), see the clip from the geological map showing quaternary deposits in Figure 13. Figure 13: Map of Quaternary deposits with the location of observation well (star) and area for site investigations (ellipse) marked (SGU, n.d.-g). © Sveriges Geologiska Undersökning. West of the well, the bedrock outcrops and forms a local height. Viktor Rydbergsgatan (south of the well) follows a depress filled with postglacial sand, and east of the depression, the bedrock outcrops and forms another height. North-east of the case study area, the postglacial silt changes to postglacial clay that the whole inner city is built on (SGU, n.d.-g). South from the site (outside the small clip of the map in Figure 13), at higher terrain level, consists the ground of glacial clay. The case study site is placed at the boundary between the glacial and postglacial clay areas. The whole Gothenburg area has been below the highest shoreline (Andréasson, 2015), which affects the expected soil stratigraphy. The wave-washed gravel is sorted by waves at the edge of a sea or a lake before the land raised above the sea level. The postglacial silt and clay are sediments settled after the drawback of the ice. Both silt and clay consist of very small particles that deposit in calm water. Postglacial sediments are usually underlaid by glacial sediments, such as glacial clay. In general, the glacial sediment is underlaid by layers of glaciofluvial sediments or till. These layers vary in spreading and in thickness from only some single meter to tens of meters. A layer of glaciofluvial sediments or till forms in this situation a confined aquifer that is recharged where the material reaches the surface, often next to bedrock CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 29 outcrops, or if the layer is in contact with another water-bearing layer such as postglacial sorted sand or gravel. A site investigation has been done for other reasons approximately 50 meters south- east of well location (see Figure 13), in the grass-covered area next to Viktor Rydbergsgatan. The historical site investigation includes both geotechnical sounding with CPT (cone penetration test) and hydrogeological testing performed in installed wells (Sweco Civil AB, 2019; Sweco Environment AB, 2018). The hydrogeological testing includes slug-test and sieve-analysis on disturbed soil samples. The interpretation of the CPT-sounding shows varied soil stratigraphy with several layers of sand, silt, and clay. The layers closer to the ground seems to be more permeable (consists of more sand and silt) while the deeper layers seem to be less permeable (consists of more silt and clay). The depth to bedrock in this part is between 4 and 7 meters and the bedrock is overlaid by a till layer with a thickness of 0.5-1 meter. A well was installed within the investigation area with the filter in the till layer, and the hydraulic conductivity of this layer was evaluated with slug-test in the well. The evaluation of the slug-test gave a transmissivity of the till layer in the magnitude 1*10-6 and a storativity of 0.025. Sieve- and sedimentation analysis was performed on soil samples from around 2.5 meters depth. The hydraulic conductivity was evaluated to be in the magnitude of 10-9 to 10-8 m/s, which corresponds to fine silt and clay (Carlsson & Gustafsson, 1997). The conclusion from the findings by the site investigation are that the stratigraphy at the site is complex and not follow the expected stratigraphy due to location below the highest shoreline. The soil depth in the area around the well is about 4.5 meters according to the soil depth map from SGU (SGU, n.d.-h). The map shows an increase in soil depth east from the case study site (in areas with clay cover), and no soil at all in the area where the bedrock outcrops. The geological maps showing bedrock implies that the bedrock in the area consists of metamorphic Granitoid from the Sveconorwegian orogen (1.6-1.5 Ga) (SGU, n.d.-c, n.d.-d). The maps do not show any deformation zones (fractures in a deformation zone may be connected to the soil aquifer and can affect the groundwater levels) in the close surroundings to the case study area. 4.1.3 Hydrogeology According to the map with groundwater aquifers from SGU, it is a confined aquifer under the clay in the central parts of Gothenburg (SGU, n.d.-e). The edge of this aquifer is relatively close located to the case study site. The aquifer has an extraction potential of 1-5 l/s, see Figure 14. The water in the aquifers flows from the edges towards the middle parts. The case study area is likely a recharge area to this big confined aquifer under the clay. The extend of the local aquifer around the well is more difficult to evaluate from maps like these. The surface water runoff from the area is towards Göta Älv (SMHI, n.d.-e), and the groundwater usually flows in the same direction as surface water, so the groundwater discharge is probably towards the Göta Älv. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 30 Figure 14: Soil aquifers in central Gothenburg (SGU, n.d.-e). © Sveriges Geologiska Undersökning. 4.1.4 Hydrology The case study site in Lorensberg is in an urban environment and that affects the water movements in the area e.g. by more surface runoff and less groundwater recharge. The actual well is placed in a small park area called Himlabacken. The park's surface is covered with grass and some trees and shrubs. The near surroundings to the well may allow infiltration of water to the ground, but the park is surrounded by streets, buildings, and impermeable surfaces (e.g. parking lots paved with asphalt). The hardened covers create surface runoff that may be received by the stormwater system, runoff to surface water bodies, or infiltrate to the ground if it reaches areas where it is possible. The ground surface falls from higher elevation south-west of the park area (Kapellplatsen/Läraregatan) towards lower elevation north-east of the park area (Götaplatsen/Avenyn). It implies that surface runoff from the higher elevation south-east of the park area may end up within the park. The total amount of precipitation during a year in Gothenburg is around 800 mm (SMHI, n.d.-c) and the evapotranspiration is about half (400 mm) (SMHI, n.d.-a). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 31 4.1.5 Observation well and reference well characteristics Dimensioning groundwater levels are calculated with the equation from TK Geo as described in the method chapter earlier in the report, see section 3.2. To enable the calculations, a representative reference well for the case study site is needed. As observation well for the case study site Lorensberg, a well from the city’s groundwater program is used (Göteborgs Stad, n.d.). The well is notated GW485. Well GW485 is located in a recharge area to the large confined aquifer under the clay in the central parts of Gothenburg, but the location of the well within the local aquifer is more difficult to define. The complex stratigraphy in the area makes it hard to interpret how water infiltrates and flows in the ground. The well filter is placed in the lower aquifer (Göteborgs Stad, n.d.). The well is 4.64 meters deep and that depth corresponds to the soil depth at the site given at the soil map by SGU. This means that it is reasonable to assume that the filter is placed in the till layer according to the site investigation performed in the area. If the aquifer in the till is confined or unconfined is more difficult to evaluate. The area has a very complex soil stratigraphy and the till is overlaid by both permeable and less permeable layers. If the water can infiltrate directly to the till depends mainly on the horizontal spreading of the impermeable layers and thus if the water can find a way down. In this case, may the local aquifer be something between a confined and unconfined aquifer. As reference well, wells from SGU’s groundwater program are used. The data within SGU’s groundwater program is openly available at their website (SGU, n.d.-f) (and also at other websites that collect geodata from several authorities) and is free to use under the creative commence license CC-BY 4.0. For this site, well SGU 52_2, which is placed in Sandssjöbacka nature reserve south of Gothenburg, is used (SGU, n.d.-f). The well is selected since its characteristics are similar to the observation well (see Table 2) and it fulfils the requirements in TK Geo (within 50 km). Table 2: Characteristics of observation and reference well. Characteristics Observation well, GW485 Reference well, SGU 52- 2 Aquifer type Confined/Unconfined Unconfined Well placing in the aquifer Recharge Recharge Soil type at the surface Postglacial sand Glacial clay Soil type at the filter Till Till The observation well, as described earlier, is placed in an aquifer that is hard to classify as either confined or unconfined. The surface material (postglacial sand) allows water to infiltrate (thus unconfined) but the many different soil types with lower permeability in the stratigraphy above the till layer may affect the amount of water that reaches the aquifer in the till. The reference well (SGU 52_2) is classified as unconfined by SGU (SGU, n.d.-f), but classified as confined by Svensson & Sällfors, (1985). The soil type at the surface is glacial clay according to the geological map showing quaternary deposits (SGU, n.d.-g) and that indicated thus confined conditions. Therefore, may also this well be placed in something between confined and unconfined conditions. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 32 To further confirm that the selected reference well is suitable for the case study site is the groundwater levels during the six months long observation period plotted for both wells to see that the levels covary, see Figure 15 and Figure 16. Figure 15: Groundwater levels in the observation well for six months observation period. Figure 16: Groundwater levels in the reference well for six months observation period. As can be seen in Figure 15 and Figure 16 covaries the levels in the two wells to a high degree. Only some small differences could be seen by visual inspection, so this reference well seem to be representative of the case study site. 19 19,4 19,8 20,2 20,6 21 21,4 m .a .s .l G.w.l Observationwell 51,2 51,6 52 52,4 52,8 m .a .s .l G.w.l Referencewell CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 33 4.1.6 Dimensioning groundwater levels The results from the calculations of dimensioning groundwater levels (both maximum and minimum levels) according to the TK Geo method and with extreme value analysis are presented in Table 3 and Table 4. For this case study site, only dimensioning levels with 3- and 6-months observation period is calculated since the period with high-resolution data is too short to allow calculation with 12 months observation period. For the 3 months observation period, data from September 1, 2016, to December 1, 2016, are used and for 6 months observation period, data from September 1, 2016, to March 1, 2017, are used. This period is simply selected as the observation period since high-resolution data are available for the observation well. The high-resolution data are measured with automatic pressure loggers. For the calculation of dimensioning groundwater levels, measurements twice a month are needed (in total 7 recorded levels for three months' observation period) and that is selected from the logged data by sort out the first value of the day for the 1st and 15th of every month. Even for the reference well, the first measured value from the 1st and 15th of every month are selected to be included in the observation period series. For calculation of dimensioning levels with long return times, required in the term 𝑆𝑟 in Equation 2, the whole available measurement series is used. Table 3: Dimensioning groundwater levels, maximum levels, according to the TK Geo method (with 3- and 6-months observation period) and with e