Characterizing hydrological processes within kettle holes using stable water isotopes in the Uckermark of northern Brandenburg, Germany

Understanding the hydrologic connectivity between kettle holes and shallow groundwater, particularly in reaction to the highly variable local meteorological conditions, is of paramount importance for tracing water in a hydro(geo)logically complex landscape and thus for integrated water resource management. This article is aimed at identifying the dominant hydrological processes affecting the kettle holes' water balance and their interactions with the shallow groundwater domain in the Uckermark region, located in the north‐east of Germany. For this reason, based on the stable isotopes of oxygen (δ18O) and hydrogen (δ2H), an isotopic mass balance model was employed to compute the evaporative loss of water from the kettle holes from February to August 2017. Results demonstrated that shallow groundwater inflow may play the pivotal role in the processes taking part in the hydrology of the kettle holes in the Uckermark region. Based on the calculated evaporation/inflow (E/I) ratios, most of the kettle holes (86.7%) were ascertained to have a partially open, flow‐through‐dominated system. Moreover, we identified an inverse correlation between E/I ratios and the altitudes of the kettle holes. The same holds for electrical conductivity (EC) and the altitudes of the kettle holes. In accordance with the findings obtained from this study, a conceptual model explaining the interaction between the shallow groundwater and the kettle holes of Uckermark was developed. The model exhibited that across the highest altitudes, the recharge kettle holes are dominant, where a lower ratio of E/I and a lower EC was detected. By contrast, the lowest topographical depressions represent the discharge kettle holes, where a higher ratio of E/I and EC could be identified. The kettle holes existing in between were categorized as flow‐through kettle holes through which the recharge takes place from one side and discharge from the other side.

These ponds provide permanent habitats for numerous threatened and endangered Red List species as well as presenting important migratory habitats to waterfowl (Hildrew, Townsend, & Francis, 1984;Platen et al., 2016;Prowse & Conly, 2000). Due to their location within often agriculturally intensive regions, kettle holes are regularly subjected to large anthropogenic forcing including pollution by agrochemicals, the removal of habitats, and buffer zones with farm machinery as well as unsustainable wetland drainage practices (Kalettka & Rudat, 2006;Lischeid, Kalettka, Merz, & Steidl, 2016).
Though kettle holes are of small size, it has been shown that depressional wetlands and small water bodies <0.001 km 2 give rise to disproportionately greater atmospheric carbon emissions when compared with other, often larger surface water bodies (Holgerson & Raymond, 2016;Lischeid et al., 2018). Moreover, recent studies on similar small water bodies and lakes in Canadian freshwater ecosystems have shown prominent declines in water levels resulting from climate change (Brock et al., 2009;Emmerton, Lesack, & Marsh, 2007).
Studies of kettle holes within central European postglacial landscapes have until recently, focused predominantly on biological interactions between macrophyte plant species and zooplankton with internal water quality parameters and internal biogeochemical cycling at the local scale (Joniak et al., 2007;Joniak, Kuczynska-Kippen, & Gabka, 2017;Joniak, Nagengast, & Kuczynska-Kippen, 2009;Kuczynska-Kippen & Joniak, 2016;Patzig et al., 2012). Though the degree of biological interaction among kettle holes is an important current prominent research topic, a detailed hydrological investigation of central European depressional wetlands is presently lacking. A consideration for the hydrological conditions within and between kettle holes is essential considering the importance for understanding the impacts that climate change may have on these unique wetland regions as well the development of hydrologically controlled agricultural pollutant pathways (EC., 2006).
Many previous studies on central European kettle holes have considered them frequently as hydrologically isolated, individual water bodies located above regional groundwater. The filling dynamics were implied to be dominated chiefly by winter run-off from snowmelt and rain on frozen soil and drying by significant summer evaporation producing pronounced wet-dry cycles without significant groundwater interaction (Kalettka & Rudat, 2006;Nitzsche et al., 2017). Hydrogeological investigations however from similar age Pleistocene kettle holes within the prairie region of north America and Canada have propounded that a more significant interrelation between groundwater and kettle holes can exist (Dempster, Ellis, Wright, Stone, & Price, 2006;Hayashi, van der Kamp, & Rudolph, 1998;Winter & LaBaugh, 2003). Within the prairie region, kettle holes have been identified to form topographically dependent, shallow groundwater connected networks, allowing the delimitation and classification of kettle holes into recharge, flow-through, and discharge hydraulic types (Hayashi et al., 1998;van der Kamp & Hayashi, 2009).
In order to address the question of kettle hole hydrology and groundwater dependence, several tools are available including the installation of hydrometric devices for water level derivation coupled with modelling approaches. Although these methods are locally highly useful, they suffer from high costs and limited spatial coverage often to a few specific sites. This is particularly restricting within regions of high spatial hydrologic variability and numerous water bodies such as within Pleistocene landscapes (Yi, Brock, Falcone, Wolfe, & Edwards, 2008). Despite the fact that the skills of mechanistic hydrological models are often oversold, there is a broad agreement on their capabilities in plausible simulation and prediction (Beven, 1989;Woolhiser, 1996).
However, the validity of physics-based hydrological models for process simulation poses formidable challenge mainly due to their computational burden/overparameterization resulted from the watershed heterogeneity and challenges with model identifiability. For the given reasons, their applicability to large scales has been restricted for practical decision support role (Ratto, Castelletti, & Pagano, 2012).
Globally, the stable isotopic composition of terrestrial waters is typified by the global meteoric water line (GMWL) of Craig, Gordon, and Horibe (1963): δ 2 H = 8 × δ 18 O + 10. On a local level, however, meteoric waters often possess a variable composition producing a local meteoric water line (LMWL; Brock et al., 2009;Craig et al., 1963). Typically during progressive evaporation of a water body such as a kettle hole, isotopic fractionation leads to enrichment of heavy stable isotopes of 2 H and 18 O within the residual water whilst the lighter isotopes of 1 H and 16 O are removed to a vapour phase. The result is the systematic evolution of kettle hole isotopic composition in δ-space along a local evaporation line (LEL) whose gradient is typically smaller than both the GMWL and LMWL. The relative position and movement of the isotopic composition of a water body along the LEL have been shown to relate intimately to the water balance of that specific water body, which may be quantified using isotope mass balance models (Brock et al., 2009;Gat & Gonfiantini, 1981;Gibson & Edwards, 2002). The outputs from these isotope mass balance models have been proposed to represent an effective method by which to classify water bodies into groups based on the degree of hydrologic connectivity or isolation from streams and/or groundwater (Brock et al., 2009). So far, no attempts have been made to classify central European kettle holes on this basis.
Kettle holes have a high informative value functioning as a suitable indicator for changes of regionally connected hydrological system. The unsolved challenge of kettle holes arise from their complexity, high abundance, and a stunning degree of heterogeneity to a wide range of scales and parameters including media properties (e.g., hydraulic conductivity), fluxes (e.g., run-off), or state variables (e.g., soil moisture) that necessitate advancing our understanding of their elaborate hydrological system (Brock et al., 2009;Skrzypek et al., 2015). The challenge becomes even more formidable for the characterization and quantification of complex and sporadic interaction between seemingly isolated kettle holes and their adjacent shallow groundwater as a result of high dynamics of the effective transmission zone where the flux exchange takes place between the two domains (e.g., groundwater and kettle hole; Brannen, Spence, & Ireson, 2015). This quantification is important not only for water resource management but also for maintaining the biodiversity of kettle holes in order to resolve conflicts related to water use and for restoring water ecosystems.
Several studies have sought to characterize and distinguish the water balance components that possibly contribute to kettle holes located across central North America (Haque, Ali, & Badiou, 2018;Hayashi, van der Kamp, & Rosenberry, 2016;Hayashi et al., 1998;Hayashi, van der Kamp, & Schmidt, 2003;Lissey, 1971;Neff & Rosenberry, 2018;Upadhyay, Pruski, Kaleita, & Soupir, 2019;van der Kamp & Hayashi, 2009) that encompass three southern Canadian provinces and five upper Midwest states, that is, the prairie pothole region. Nevertheless, by contrast, the importance of kettle holegroundwater interactions in the Uckermark region has remained as one of the most important questions underlying the investigations into the hydrology of these small water bodies. To that respect, the initial studies conducted in the Uckermark region suggested that kettle holes are disconnected from groundwater domain and they should be treated as isolated water standings (Kalettka, Rudat, & Quast, 2001). Recent studies have suggested that the isolated kettle holes can be variably connected to each other as well as to the shallow groundwater and may show seasonally variable interactions (Gerke, Koszinski, Kalettka, & Sommer, 2010;Kalettka & Rudat, 2006;Lischeid et al., 2017;Nitzsche et al., 2017). This connection could potentially be related to the topographical position of a kettle hole with respect to the shallow groundwater system (Lischeid et al., 2018).
With respect to the major water suppliers of potholes, pond's permanence of the prairie pothole region is highly dependent on the direct rainfall on the potholes and upland run-off generated form snow drift, snowmelt run-off, and occasional summer run-off during heavy rains (Brooks et al., 2018;Hayashi et al., 2016). By contrast, the pond's permanence of the kettle holes in the north-east of Germany is heavily reliant upon the shallow groundwater inflows Lischeid et al., 2018;Nitzsche et al., 2017) and is partially supplied with the direct rainfall on ponds and diffuse run-off produced as a result of rainfall occurrence on the frozen soils (Gerke et al., 2010). Therefore, as there are a few studies in the north-east of Germany (where the kettle holes are highly scatted) that have attempted to gain insights into the water storage suppliers of these kettle holes, the results of this study can be compared with an intensive and progressive research undertaken so far in the North of America. Consequently, we can enhance our understanding of the effects of geological setting, meteorological condition, and dominant hydrological processes that sustain the pond permanence (hydroperiod) in each of these distinguishable systems.
In this study, first, we aim to identify the main factors controlling the observed patterns of enrichment and dilution phases in kettle holes by investigating the isotopic composition of samples taken from rainfall, 50 kettle holes across the Uckermark region of north-east Brandenburg, and observation wells installed in their nearby. Second, we explore the possible connectivity of kettle holes to shallow groundwater based on the calculation of evaporative losses across the region via evaporation/inflow (E/I) ratios obtained from the isotopic mass balance model "Hydrocalculator." Third, we investigate the topographical-driven connectivity among the kettle holes by analysing the relationships between E/I ratios and electrical conductivity of kettle holes with their respective landscape elevations. Based upon these results, a hypothetical landscape model for the Uckermark region portraying hydrologic connectivity among the kettle holes and in relation to their adjacent shallow groundwater domain will be proposed.
Addressing the research questions set for this study provides profound insights into tracing and quantifying the waters through the hydrologic cycle, which has not been taken into consideration for the kettle holes located in the north-east of Germany in relation to whose adjacent shallow groundwater system. We anticipate that our findings will be of use in future water management decisions to protect these wetlands, particularly as farmers expand crop production in this region.  (Merz & Steidl, 2015). Potential evapotranspiration, calculated as reference grass evapotranspiration, according to Allan et al. (1998) ranges from 570 to 635 mm/year. Annual groundwater recharge, on average, is rather low and changes from 70 to 90 mm/year (Lahmer & Pfützner, 2003). Long-term mean discharge 1972-1990 of the Quillow stream after confluence with the stream from the adjacent catchment amounted to 143 mm/year . Despite that fact that the region is characterized by an abundance of F I G U R E 1 Overview of the geographical location of the Uckermark region in relation to Germany, as well as the location of all kettle holes included in the 2017 study. Red triangles correspond to individual kettle holes sampled during the 2017 sampling campaign F I G U R E 2 Meteorological variables recorded at Dedelow meteorological station over the period of January 2015 to August 2017 groundwater resources, wetlands, and in particular kettle holes, major fraction of the precipitation evaporates, and surface run-off is negligibly small (Nutzmann & Mey, 2007), generating mostly from agricultural land (Schindler, Müller, Thiere, & Steidl, 2004).
The region demonstrates a moderately undulating topography between 40 and 112 m above sea level (MASL), typical of hummocky ground moraine landscapes with abundant kettle holes and wetlands occurring between hummocks. The number of classified kettle holes in the Uckermark region has been estimated to be greater than 1,550 based on extensive fieldwork over the past decades (Kalettka & Rudat, 2006;Nitzsche et al., 2017). Geologically, the region itself consists of Pleistocene glacial sediments derived from three large glaciations of the Elsterian, Saalian, and Weichselian. The present landscape of the Uckermark region was particularly strongly shaped during the Weichselian glaciation (Ehlers, Gibbard, & Hughes, 2011). The regional groundwater structure of Brandenburg is closely related to the Pleistocene glacial landscape, with different regional groundwater systems being delineated on the basis of large geomorphological structures including end moraines, glacial valleys, and till upland regions. Sediment types deposited over the study area consist of a range of clastic sediments of glacial fluvial origin and till. Typically, permeable and porous sands form aquifer bodies, and more impermeable and lower conductivity till beds consisting of clay, silt, and calcium-containing sands forming predominantly aquitards (Ehlers, Grube, Stephan, & Wansa, 2011;Merz & Pekdeger, 2011;Merz, Steidl, & Dannowski, 2009).
Altitude decreases from 80 MASL in the western part of the catchment to 30 MASL in the south-east (Glacial Valley of the Ucker). Correspondingly, regional groundwater flow is directed to the east/ south-east; the River Quillow is the main drainage recipient of the region. Hydraulically, these regions are characterized by regional transient and local recharge dynamics of the groundwater. Local, unconfined, and only temporarily saturated aquifers of Weichselian age are deposited above the till layers. Groundwater is suggested to be recharged in these upland areas and discharged into lowland wetlands, lakes, kettle holes, and river systems. Hydraulically, these local shallow aquifers are connected relatively quickly to the smaller surface water system; thus, a height dependence on recharge and discharge dynamics has been suggested (Merz & Pekdeger, 2011).

| Selection of kettle holes
Fifty kettle holes were selected for this study on the basis of predetermined biogeochemical and ecology importance in the Uckermark region (Tables 1 and 2; Figure 1). The selection was based on previous work by Kalettka and Rudat (2006) who presented a classification of kettle hole. Hydrogeomorphic types delineated from various physical and biogeochemical parameters. The selected kettle holes were chosen to cover a variety of Hydrogeomorphic types (Table 1), vegetational succession stages, and landscape elevations (Table 2). In order to increase spatial coverage and sample size, seven more kettle holes were later included into the study from all vegetation succession stage types. During sample periods, sampling was carried out monthly with a column bucket sampler in order to minimize the disturbance of the water column. It is lowered into each kettle hole body to a depth of 10 cm below the water surface near the kettle hole centre region following literature suggested sampling practices (Brock et al., 2009;Yi et al., 2008). However, this approach was not always possible due to high water depths and soft, unpassable sediments during the early sampling months, as well as vegetation growth and obstruction later within the summer. Where the centre region was unreachable, a suitable alternative edge proximal location was selected and consistently sampled each month. In order to justify this approach, isotopic samples were taken from the edge and centre of two, well studied, edgetype kettle holes (259 and 807; Table 1) during the May sampling period and compared (Table 3). Due to shallow system characteristic of the studied kettle holes, represented by a water depth less than 150 cm in nearly all of the kettle holes, and assuming a well-mixed setting, we did not measure the temperature profiling for the kettle holes, whereas we provided the isotopic composition profiling.

| Sampling strategy and data detection
Water samples obtained for isotopic analysis were collected from each waterbody and were immediately transferred to high-density polyethylene sample bottles, filled, and sealed to minimize evaporative loss of water during further processing and transport.
In addition to surface water samples taken from the kettle holes, water samples for isotopic analysis were obtained from four groundwater wells surrounding both kettle holes 807 and 259 using a battery-operated groundwater pump. Prior to groundwater abstraction, each groundwater well was pumped until gone dry and allowed to recharge, thus ensuring that solely groundwater was sampled and that no prior in-well mixing with external sources had occurred.
Four shallow monitoring wells are installed in the close vicinity of kettle holes 807 and 259 to monitor groundwater levels at the shallow unconfined layer. The observation wells are inserted into a depth of 2 to 4 m, out of which 1 m at their bottom is covered by screen.
Depth to the groundwater level in the vicinity of kettle holes 807 and 259 varies from1 to 2 m, depending on the growing season and precipitation events.
T A B L E 1 Kettle holes sampled during 2017 based on vegetational succession stage and hydro-geomorphology as defined by Kalettka and Rudat (2006) 2.3 | Calculation of kettle hole E/I ratios Based on the assumption of connectivity of kettle holes with shallow unconfined groundwater, E/I calculations of this study assumed a steady-state approach. This assumed that kettle holes were continually replenished by inflowing shallow groundwater water that compensated for evaporative loss and groundwater outflow (Skrzypek et al., 2015). In other words, the water level in the kettle holes are a subdued replica of that of the adjoining shallow groundwater (Lischeid et al., 2018). As a result, they fluctuate in unison.
Although due to the recharge flux resulted from precipitation, the groundwater level rises quicker rather than that of a kettle hole due to the porosity-driven effect of porous material, they will reach to an equilibrium ultimately. Assuming the groundwater allowed adequate replenishment of water losses to kettle holes, the E/I ratio was calculated using the Hydrocalculator programme for each individual kettle hole using a reformulated equation (Allison & Leaney, 1982;Mayr et al., 2007;Equation 2). The required input variables are listed in Table 4.
where δ P shows initial value of isotopic composition of kettle hole surface water and δ L is the isotopic composition of kettle hole water at the subsequent sampling interval, δ * states the limiting isotope enrichment factor (Equation 3), and m is the enrichment slope (Equation 4).
δ * is normally computed by means of air humidity (h), the isotope composition of moisture in ambient air (δ A ), and a total enrichment where ε states the total fractionation factor (Equation 5) that is a summation of the equilibrium isotope fractionation factor (ε + ; Equation 6) and the kinetic isotope fractionation factor ε k (Equation 7; Gibson & Reid, 2010). Note: Some kettle holes were excluded from the DEM height classification due to a lack of water present at any time during the study period. According to a Hydrogeomorphic classification made for the kettle hole of the north east of Germany by Kalettka and Rudat (2006), puddle-type kettle holes are "very small and characterized by a nonpermanent shore due to periodic complete use as arable land, especially in dry periods." a Kettle holes with catchment size ≥31.6 ha (Kalettka & Rudat, 2006 where C k represents the kinetic fractionation constant, which is  Gonfiantini, 1986).
Following Horita and Wesolowski (1994), α + is calculated for hydrogen and oxygen isotopic compositions using Equations (8) and (9) where temperature (T) is given in Kelvin degrees.
The initial value of input water (δ P ) was decided as the isotopic composition of kettle hole surface water during February when the evaporation is minimum and as a result of the precipitation events, the maximum dilution can occur, irrespective of their setting (flowthrough or discharge). Therefore, in this period (dilution phase), the maximum water depths were also expected ( Figures S1-S2). In cases where isotopic depletion occurred in kettle holes from February to March, filling was assumed to have been completed in March, and this value utilized for the input (δ P ). The outflow from each kettle hole (δ L ) was taken as the isotopic composition of kettle hole water at the subsequent sampling interval, with thus the final δ L being obtained in August 2017. The isotopic composition of atmospheric moisture δ A was difficult to measure in the field and was thus calculated using the isotopic composition of precipitation between sampling periods (δ Rain ), which was computed with respect to stable isotope composition of precipitation, equilibrium isotope fractionation factor, and the slope of LEL. The individual kettle hole specific LEL was derived from regression of all isotopic samples obtained at each analysed kettle hole in 2017, which were then included in Hydrocalculator ( Figure S4). δ Rain was obtained from monthly regional measurements of the isotopic  2.4 | Monthly regional LELs and regional evaporative loss calculation In addition to the calculation of E/I ratios, it has been suggested that to estimate the water balance changes at a regional scale, differences in gradients of the regional LEL in monthly intervals may be used (Nitzsche et al., 2017). Regional monthly LELs were constructed from regression of all isotopic samples obtained in each sampling month ( Figures S5-S7). This methodology implied firstly calculating the difference in LEL gradients between two monthly sampling intervals from the LMWL. Second, these normalized LELs were then 3.2 | Isotopic framework for evaluating kettle hole water balances 3.2.1 | LELs, local meteoric water line, and regional water loss calculation In this case, the isotopic composition of several kettle holes in July and August lay close to the isotopic composition of summer precipitation at the LMWL Figure 4b.
Regional water loss calculated using monthly regional LELs from February to July 2017 was calculated as 27.8%. This water loss did not occur uniformly with the greatest loss reported from April to May (21.2%). Furthermore, two periods of water gain were observed.
These gain phases were witnessed as a slight gain from February to March (2.3%) and a more significant gain of 15% from July to August.

| E/I ratios for February-August 2017 period
Generally, E/I ratios in the earliest months (February to March) of the February-August period clustered close to zero with a mean E/I ratio across all kettle holes during this period of 0.01 (1% evaporation of inflow; Figure 5b). The variance about the mean during this period was very low (0.0004). The wood-type kettle hole (040)     F I G U R E 9 Correlation between electrical conductivity ratios for all kettle holes with their respective elevations in 2017 accordingly resulted in negative E/I values ( Figure 5). On appraisal of the regional LEL for June and July, it may be seen that the most enriched kettle hole isotopic values leading to the greatest axial shift along the LEL with regard to both δ 2 H and δ 18 O occurred during June 2017. It has been widely discussed that a lowering of the LEL gradient over the course of the summer enrichment phase is also indicative of isotopic enrichment due to local environmental and climatic conditions acting to reduce the LEL slope (Skrzypek et al., 2015). The lowest LEL gradient for 2017 (see Figures S5-S7) was observed for kettle holes during July 2017 (3.90; Figure S7a) after which a stark increase in gradient was observed in August ( Figure S7b). Moreover, the calculated mean E/I ratios across all kettle holes reached its maximum from the May-June period before decreasing rapidly to negative E/I ratios in the June-July period (Figure 7b). To that respect, Nitzsche et al. shallow groundwater flow (Hayashi et al., 1998;van der Kamp & Hayashi, 2009;Winter & LaBaugh, 2003).

| Correlation between altitudes of kettle holes and their electrical conductivity
In comparison with the previous studies conducted in the northeast of Germany (e.g., Kalettka et al., 2001) as reported that the topographically isolated kettle holes are not hydrologically connected, our findings provide sufficient evidence supporting the role of the hydrologic connectivity among the seemingly isolated kettle holes. Nevertheless, the term "isolated" has been criticized as a misnomer (Mushet et al., 2015), because physiographically isolated kettle holes can yet influence the watershed integrity. Groundwater connectivity is the only possible type of hydrologic connection between isolated kettle holes and downstream water bodies in areas that drain internally and have no surface water outlet, as was found to be the case in Uckermark. was caused by a larger monsoon precipitation amount and hence groundwater inflow than was observed in successive years of higher E/I ratios (Cui et al., 2017).

|
A simple water balance for each sample period using lysimeterderived evapotranspiration data and precipitation amounts showed that following a generally negative water balance from March to June 2017, within the subsequent June to August period, the water balance became almost balanced ( Figure 10). It has been widely discussed that the Brandenburg region as a whole demonstrates a negative water balance throughout the summer months with high annual evapotranspiration rates (Merz & Pekdeger, 2011). The meteorological data obtained for 2017 however shows that it is possible to have a reduction in the negative water balance and even a positive summer water balance under exceptional rainfall conditions such as that witnessed in June and July 2017, which can lead to substantial refilling of kettle holes during these months. Having linked this to the isotopic compositions, a similarity can be drawn between the positive and negative water balance periods with the isotopic compositions during the enrichment and dilution phases (Figure 3c,d).
4.5 | Calculated regional water loss from Uckermark kettle holes Mean E/I ratios calculated from all 50 kettle holes in 2017 suggested average regional water losses of 26% (E/I = 0.26). In other words, 26% of total inflowing water to all of the kettle holes is lost each year. Assuming no inflow to the kettle holes, all of their respective storage can be lost during approximately 3.5 years (0.26 × 3.5 ≈ 1). Percentage enrichments suggested an average isotopic enrichment during kettle hole flow-through of 28.10%. Thus, the regional mean water loss during the 2017 February-August period was likely to be approximately 26.00-28.10%. Nitzsche et al. (2017) using the difference in LEL between a defined dormant and growth season in a similar method to the monthly LEL-based approach in present study proposed a regional water loss of 28.00% in both 2013 and 2014, advocating the findings of this study. Moreover, Hayashi et al. (1998) found that summer evapotranspirative losses in Canadian prairie wetlands approximated 25.00% based on pan experiments, with a significant (75.00%) loss through infiltration and lateral, shallow groundwater flow to the sides of kettle holes in close agreement to findings of this study and again supportive of substantial lateral groundwater flow-through (Hayashi et al., 1998).
The regional connection was already discussed by Kayler et al. (2018) ranged to values as low as 1.6 years (Anderson & Cheng, 1993 8 and 9). Student t tests calculated for elevation classes based on a one-tailed approach produced statistically significant results with p values <.05. For classes based on electrical conductivity, a p value of .0012 was calculated. A p value value of .026 was found also for E/I ratio-based elevation classes (Figure 9).
Based on total inflow to evaporation ( ITOT = E ) calculated by an isotopic mass balance model by Isokangas, Rozanski, Rossi, Ronkanen, and Klove (2015), a poor connection between groundwater and 67 subpolar lakes in Finland was detected, although deeper flow paths from high-laying lakes towards low-laying lakes were characterized. Thus, the effect of lake's altitudes on establishing of subsurface interconnection could be identified.
The increase in E/I ratios and electrical conductivity among kettle holes with decreasing elevation may suggest a general trend to water loss-dominated and groundwater discharge behaviours. Heagle, Hayashi, and van der Kamp (2013) and Hayashi et al. (1998) found similar patterns in chloride for the kettle holes in the northern prairie region of North America. They found that kettle holes and wetlands located at higher topographic elevations tended to display low solute (particularly chloride) and thus salinity. They argued that this was due to kettle holes being located above the groundwater potential causing the shallow groundwater output of solutes from the kettle hole in a recharge-type scenario. Wetlands that were observed at lower topographic elevations on the other hand were shown to have higher solute and thus salinity contents. These were suggested to be influenced by inflowing solutes from higher kettle holes that accumulated in a shallow groundwater discharge-type environment (Hayashi et al., 1998;Heagle et al., 2013). Similar results were found for small mountain lakes in the Colorado alpine and in the forest land of Wisconsin whereby electrical conductivity was observed in both cases to be higher at lower elevations. Increases in electrical conductivity were hypothesized to result additionally from shallow groundwater flowthrough entraining increased solutes from weathered material with decreasing height (Barsch & Caine, 1984;Kantrud, Krapu, & Swanson, 1989;Swanson, Caine, Woodmansee, & Kratz, 1988).
Using the results obtained within this study and findings from kettle holes located in the prairie region, a simplified conceptual model of Uckermark kettle holes is illustrated by Figure 11. This model is based upon the dominance of shallow groundwater flowthrough from higher elevations to lower elevations, leading to increased solute concentrations and E/I ratios.
According to the proposed conceptual model (Figure 11), dissolved solutes are infiltrated into groundwater in higher kettle holes via flow-through with partial recharge and transferred via shallow groundwater flow-through to kettle holes occupying lower landscape elevations and hence concentrated in them by partial discharge. The increase in E/I ratio further supports this by suggesting a height transition from partially open, flow-through dominated to increasingly partially closed, water loss-dominated kettle holes with reduced height.
This implies that kettle holes though predominantly flow-through dominated may show admixtures of recharge or discharge behaviours depending on landscape elevation. More work within this area is needed to be able to confirm this behaviour.

| CONCLUSION
The current study has been conducted with the purpose of pinpointing the hydrological processes that influence kettle hole water balance from February to August 2017 in the Uckermark of northern Brandenburg, Germany, where kettle holes are densely populated.
The present investigation has taken advantages of stable isotopes of oxygen (δ 18 O) and hydrogen (δ 2 H), and deuterium excess to shed lights on complex and sporadic interactions between kettle holes and the groundwater system. To that end, isotope mass balance model implemented in the "Hydrocalculator" programme was employed.
Results indicated that isotopic composition of kettle holes has fluctuated both spatially and temporally, thereby representing distinct phases of dilution and enrichment being attributed to hydrological inflows of shallow groundwater, snow, rainfall, and evaporative loss.
Moreover, findings revealed that evaporation/inflow (E/I) ratios changed annually, which may arise from the impacts of meteorological conditions on the depth of shallow groundwater. We identified that puddle-type kettle holes exhibited the highest E/I in comparison with the other kinds of the studied kettle holes. This might be associated with intimately interlocked linkage between puddle-type kettle holes and the shallow groundwater system, which per se is highly dependent upon the precipitation in terms of wet and dry spell periods. The obtained results demonstrated that water loss over Uckermark in 2017 can change from 26.00% to 28.10%. Moreover, based on the obtained results of this study and the system understanding gained from the applied approaches to kettle holes of the prairie region, a simplified but informative conceptual model was developed. The model illustrates that across the highest altitudes, the recharging kettle holes are dominant, where a lower ratio of E/I as well as a lower electrical conductivity is seen. Conversely, the lowest topographical depressions represent the discharge kettle holes, where a higher ratio of E/I and electrical conductivity can be observed. The kettle holes that exist in between, that is, the highest and lowest altitudes, are categorized as flow-through kettle holes in which the recharge takes place from one side and discharge from the other, depending upon the regional groundwater head gradient.
Based on the findings of this study, compared with the hillslope run-off and local meteorological conditions represented mainly by precipitation and evaporation, the regional shallow groundwater system may play the pivotal role in the processes taking part in the hydrology of the kettle holes in the Uckermark region. Ergo, to quantify and characterize these reciprocal interactions between kettle holes and shallow regional groundwater system, developing a highresolution physically based hydrological model can be a viable approach; this objective should be taken into consideration for perspective studies for this region.
Therefore, previously conducted works in the field of kettle hole hydrology, which have been mainly undertaken in prairie pothole region in North America, could build our underlying assumption concerning the influence of landscape position on the hydrologic connectivity of kettle holes per se and their interactions with their respective adjoining shallow groundwater system. Our results showed that this assumption can be true for these kettle holes as well; however, further investigations are required using high resolution fully integrated mechanistic hydrological models to better quantify and characterize these hydrological processes (i.e., kettle hole-groundwater interactions) at a local scale. Due to a high abundance of kettle holes and high heterogeneity of landscape, the characterized processes obtained at a local scale should be upscaled to regional/landscape scale.
acting on behalf of the State Office of Environment, Health and Consumer Protection Brandenburg.