The Range of Ice Thrusts and Ice Piles as Reflected by Ice Scars on Trees Growing on the Shores of Coastal Lagoons: The Case of the Szczecin Lagoon

The dynamics of ice phenomena in the coastal zone, especially ice thrusting and ice piling, may be reflected by ice scars on trees. These represent damage to the lower parts of the trunks of trees growing along the shore. Such damage usually arises following the disintegration of fast ice and the resultant ice drift.

To date, tree ice scar-based studies on the dynamics of ice phenomena were performed mainly on rivers (cf. Smith, Reynolds 1983, Lederer, Garver 1996, 2000, Engström et al. 2011, Uunila, Church 2014, Pawłowski 2019). Tree ice scars are observed the most frequently along river banks, due to the permanent occurrence of a current in the river and the presence of trees growing near the waterline (Cyberski et al. 2006). On marine shores, especially along the southern Baltic coast, there are a wide number of beaches which act against ice (ice floe, sheet ice), advancing far inland. Further, a grease ice ridge forms earlier along a marine shoreline, which also acts as an obstacle to ice floe advancement (Girjatowicz, Łabuz 2020). The effects of tree damage by floe thrusting on rivers become the clearest during an ice jam. Pressured floe blocks the discharge of water, causing a considerable water level rise in the river (Beltaos et al. 2006, Lind et al. 2014, Vandermause et al. 2021). At the peak of the high water event, floe advances up the bank and may damage tree trunks. The damaged (abraded) tree trunks may reach several meters above the average water level in the river. This is evidenced by ice scars on trees growing along rivers in North America (Smith, Reynolds 1983, Lederer, Garver 1996, 2000), Canada (Uunila, Church 2014) and Alaska (Vandermause et al. 2021), and in Europe, for instance along the lower reaches of the Wisła River (Grześ 2005, Pawłowski 2005, 2019). For instance, during the ice jam on the Mohawk River in 1996, ice scars were formed on tree trunks 5 m above the normal water level (Lederer, Garver 1996). It was observed that the water level during the surge was matched by the ordinate of the maximum elevation of an ice scar on a tree trunk.

Similar damage to trees, induced by ice being thrusted on land, was observed also along the shores of coastal lagoons, straits and marine gulfs, retention reservoirs or on large lakes. Ice scars on trees are observed less frequently on these basins, however, than along rivers. In the case of lagoons, a strong wind is necessary to cause ice floe thrusting onto land. Studies on how the scale of ice floe advances onto lake shores based on tree ice scars were performed mostly in North America, Canada and the USA, where such phenomena are common due to the high number of very large lakes (Gilbert, Glew 1986, Futter 2003, Duguay et al. 2006). Sheets of ice being pressured and thrusted inland had caused large damage in the forests around the Pärnu Bay (Estonia; Orviku et al. 2011). Considerable damage had also been caused along the shores of the Luodonselkä (Finland), where grey alder trees were strongly damaged, and young Scots pines were totally destroyed (Alestalo, Häikiö 1976). Frequent cases of trees that were damaged or ploughed out together with the ground by advancing sheets of ice were also recorded in the Szczecin Lagoon area (Banzhaf 1931, Girjatowicz 2004). Similar tree damage and destruction were also observed along the shores of the Curonian Lagoon (Lundbeck 1931) and the Vistula Lagoon (Girjatowicz 2015). On the shores of the Vistula Lagoon, ice scars occurred on tree trunks not only at the ground level but also at an elevation of up to 5.5 m above the water level. Trees damaged by sheet ice thrusting were also observed around large coastal lakes of the southern Baltic region, including Łebsko and Dąbie lakes (Girjatowicz 2017), or on the shores of retention reservoirs (Jaguś, Rzętała 2000).

The elevation of the identified ice scars above the ground may be a measure of river ice jam and an ice-thrusting threat onto the bank (Beltaos et al. 2006, Pawłowski 2011, Lind et al. 2014). Ice thrusting on land causes bank erosion, damages the substratum and vegetation (including trees) and poses a threat to the infrastructure such as ports, beacons, buildings and roads and may contribute to flooding along river banks (Beltaos et al. 2006, Orviku et al. 2011, Kolerski et al. 2019).

To date, no precise information was available on ice scars on trees growing along the shores of the southern Baltic coastal lagoons, which could be used to determine the scale of the threat. The first partial study of tree ice scars was performed on the Polish part of the Szczecin Lagoon (Girjatowicz 2017). Examining the hydrological conditions of the Szczecin Lagoon is especially important as this particular basin plays an important part in transport, fishing and tourism. Numerous small ports are situated along the shores of the Szczecin Lagoon. Due to its attractive location, there is an increasing number of private investments, including houses, marinas and recreational infrastructure. All these facilities are situated on a ground that is barely raised above water, at the shore or in its immediate surroundings. The construction works are performed – not entirely reasonably – down to the shoreline, and therefore, estimating the threat degree is highly relevant, as it may enable the proprietors to avoid increased spending on protective measures in the future.

The aim of the present work is to establish which part of the coast is the most prone to the destructive impact of the ice, based on the distribution of ice scars on trees growing along Szczecin Lagoon shores. Based on the spatial analysis of the occurrence of ice-scarred trees, it was determined which shores are the most threatened by the destructive impact of sheet ice thrusting. It also obtained information on how far inland ice thrusting may reach and pose a threat. Further, the density of occurrence, distance from shore and the elevation of ice scars on tree trunks provided information on the intensity of ice impact in those regions where ice thrusting and onshore piling is the most frequent.

The present study covered the entire shore of the Szczecin Lagoon, named as Oder Lagoon (Fig. 1) in Germany, which comprised the Great Szczecin Lagoon, located entirely within Polish borders, and the Small Szczecin Lagoon, located on the German part of the border (Fig. 1B). The Szczecin Lagoon is a shallow basin (average depth: 3.8 m) connected with the Pomeranian Bay of the Baltic Sea via three narrow and very shallow straits (Figs 1A and 1B). It occupies an area of 687 km², with a volume of 2.5 km³, and its shoreline is 243 km long (Majewski 1980). Due to its very shallow depth combined with a relatively large area, the Szczecin Lagoon is characterised by a high lake exposure index value, defined as a ratio of the basin area to its average depth, equal to 191 km² m^-1 (Girjatowicz, Świątek 2021). This results in a rapid rate of water cooling (in reaction to air temperature change) and the formation of ice phenomena. The greatest natural depth of the Lagoon is only 8.5 m, and the fairway crossing the Lagoon is 12.5 m deep.

Fig. 1.

The average annual influx of fluvial waters equals 18 km³ (Radziejewska, Schernewski 2008). Marine water influx is not known due to the bidirectional exchange of waters with the sea (depending on the wind direction, water flows from the Szczecin Lagoon to the Pomeranian Bay or the opposite way). It is estimated that about 69% of water discharge from the Lagoon to the Baltic Sea occurs via Świna and its branches, 17% via Peenestrom and 14% via Dziwna (Mohrholtz, Lass 1998). As the Baltic Sea has very low salinity (on average 7.5‰ in the Pomeranian Bay; Leppäranta, Myrberg 2009) and the Szczecin Lagoon is connected to the sea via narrow and shallow straits, the lagoon is virtually a freshwater basin, with salinity ranging from 0.3 to 4.5‰ (on average, 1.4‰). The salinity is the highest in winter, when strong landward winds occur at the southern end of the Piast Canal (an artificial branch of the Świna and the main connection between the lagoon and the Pomeranian Bay). The lowest salinity occurs next to the Odra River mouth in summer (Bangel et al. 2004). The very low salinity of the lagoon enables the process of freezing of its surface waters to resemble those observed on freshwater lakes.

The mean annual water temperature in the lagoon is ca. 11°C (Schernewski et al. 2006), and the mean annual air temperature is 8.7°C, with a multiannual mean monthly air temperature reaching a minimum of 0.8°C and a maximum of 17.9°C. The mean annual precipitation sum equals 550 mm (Koźmiński et al. 2004, Radziejewska, Schernewski 2008). Mean wind velocities over the study area are the highest in winter (especially in February) and early spring (Świątek 2012), which favours the formation of ice hummocks, and thrusting ice onshore during periods of air temperature increase within this time. On the Szczecin Lagoon, ice occurs on average for 51 days. The earliest ice phenomena appear on average on 25th December (in individual years ranging from 13 November to 27 February), and the last ice in a given season disappears on average on 5th March (ranging from 13 January to 10 April), average annual maximum ice thickness equals 17 cm and the peak value reaches 50 cm (Girjatowicz, Świątek 2020).

Winds blowing from the directions ranging from the south to the west occur for 58% of the time (Fig. 1C). Wind velocity ranging from 3 m·s^-1 to 5 m·s^-1 occurs for about 50% of the time. Strong winds (with velocities exceeding 10 m·s^-1) are basically those blowing from the southern and western sectors; they represent more than 90% of the strong winds (Koźmiński et al. 2004, Łabuz 2022). Damage caused by ice floes is also impacted by water levels. On the Szczecin Lagoon, the highest water levels are observed when storm winds are blowing from the W, NW and N. Maximum water levels, recorded during winter high-water events in January 2002, 2005 or 2007 reached 0.65–0.95 m above the average water level for the southern lagoon coast in Trzebież (Kowalewska-Kalkowska, Kowalewski 2005, 2008).

Measurements, along with photographic documentation, were performed by the authors during on-foot reconnaissance campaigns along the entire lagoon. Due to the large geographic extent of the study area, fieldwork was carried out in various years in the period from 2017 to 2022. In order to facilitate the analysis and interpretation of the ice scar distribution, the Szczecin Lagoon shore was divided into six segments characterised by approximately uniform exposure: western, south-western, south-eastern, eastern, north-eastern and north-western (Fig. 1D).

Field measurements involved five key parameters of damage to trees:

Additionally, the tree species were noted, along with the tree circumference, information on fallen trees bearing scars and the scale of damage to the shore and vegetation caused by ice thrusting onshore. Geographic coordinates and the exposure of ice scars on the examined tree trunks were also noted. The aim was to determine the hydrometric conditions under which the ice scars were probably formed. The documentation was supplemented by photographs presenting the scale of the examined phenomena.

During the field campaigns, it was recorded that 147 ice scars on growing trees were distributed around the Szczecin Lagoon. Ice scars were found also on numerous trees previously uprooted by ice or due to shore erosion. Such damage was not considered in the present study, as the displaced and fallen trees did not indicate the maximum ice scar elevation, nor the distance from shore.

The distance of the fallen tree from the shore and the maximum elevation of an ice scar on a given tree were deemed the most important parameters of damage to trees. For the entire shore of the Szczecin Lagoon and for each of the segments distinguished herein, the values of these parameters were characterised in detail by means of descriptive statistics. The first stage involved a verification of the similarity of the distribution of the empirical data to normal distribution. This was achieved using the non-parametric Shapiro–Wilk normality test (Kaptein, van den Heuvel 2022). Next, central tendency measures were determined (classical – arithmetic mean and positional – median) as well as values of the lower (1) and upper (3) quartile. In order to determine whether the damage to trees in individual segments of the lagoon shore occurred at a different than average distance from the shore or were located at a different than average elevation, a difference between means test for independent samples was applied (Student’s t-test). The null hypotheses for this test were no significant differences among average distances of damaged trees from the shore or among maximum elevations of scars on trees. Those test results (both Shapiro–Wilk test and Student’s t-test) that yielded the standard significance level of 0.05 were deemed statistically significant. As part of this study, the dimensions (lengths and maximum widths) of ice scars from individual segments of the lagoon shore were compared. Further, it was attempted to determine the density of ice scar occurrence along the Szczecin Lagoon shore, following the scale proposed by Lederer and Garver (2000). These authors put forward a six-degree scale of ice scar distribution density, as follows: 0 – no ice scars; 1 – minor, exceedingly rare damage; 2 – minor, rare damage; 3 – moderate, several damaged trees; 4 – large, widespread damage, a common occurrence of scars and 5 – extreme, most trees damaged, broken trees present.

On the southern Baltic shores, the wide beaches and the presence of grease ice ridge prevent ice floes from advancing on land (cf. Girjatowicz, Łabuz 2020). The advancing ice floe crumbles and piles before the coastal grease ice ridge, and thus, the trees growing on the dunes are protected from damage. On the Baltic Sea shores, ice scars on trees are observed only over some segments of the shores of the Gulf of Riga and the Gulf of Finland (Orviku et al. 2011). On the shores of the coastal lagoons, trees growing adjacent to the shore are relatively highly exposed to damage from advancing ice floe. Following the disintegration of fast ice cover, sheets of ice readily undergo thrusting onto flat shores (cf. Kraus 1930, Dionne 1979, Gatto 1984, Lemay, Bégin 2012, Girjatowicz 2017). Trees growing along river banks are the most vulnerable to destructive ice activity. Ice floe carried by the river and ice jams forming on rivers, especially at high water levels, readily damage trees (cf. Lederer, Garver 1996, Pawłowski 2005, Uunila, Church 2014, Vandermause et al. 2021).

The average maximum ice scar elevation observed on the trees surrounding the Szczecin Lagoon equalled 1.5 m a.w.l., with a maximum elevation equal to 4.3 m a.w.l. Effects of ice thrusting represented by ice scars on trees were most frequently found on open shores or in bays free of reed or on shores with a narrow reed belt. The ice scar elevation on trees was influenced by the specifics of the hummocking process, dependent on ice type and thickness, area of the ice field or floe, wind force and water level, which facilitates ice thrusting.

Scars located high on a tree trunk are formed during the ice hummocking process (Girjatowicz 2014). As the sheet of ice thrusts up a windward slope, the ice crumbles to form brash ice on the leeward side, causing an increase in the width and height of the ice hummock. As the hummocking process progresses, the sheet of ice causes damage to tree trunks even several meters above water level. as it undergoes thrusting. Trees growing along the shore are usually an obstacle for a thrusting ice field; so, hummocking mostly occurs within such tree stands.

The largest tree damage was observed in the eastern segment of the shore (Figs 8 and 9). In this area, the damage was not only the most extensive but also the most frequent: it was documented in 47 cases, which represented nearly 1/3 of all the observed cases of damage to trees caused by ice thrusting. This is favoured predominantly by the shore exposure to the frequent western wind (Koźmiński et al. 2004, Łabuz 2022) (Fig. 1D), which has the longest fetch. Within the Skoszewo Bay, located between the Rów Peninsula and the Śmiec Peninsula (a distance of 4 km), was measured 35 ice scars on 31 trees, the most frequently, those growing close to the shoreline (Fig. 2B), including double scars characterised by different geographic exposure on four trees (Fig. 2C). Within this segment, due to shore erosion, many ice-scarred trees were already fallen or dragged onto the shore by ice and waves. The scars characterised by the highest elevation were formed in that part of the shore which is exposed to the wind from the SW direction, which favours ice run, as well as thrusting and hummocking. Wind blowing from this direction (and also from the W) is the most frequent and the strongest, especially in the winter-spring period (Koźmiński et al. 2004), causing the formation of high ice hummocks. Further, it brings warming in winter (the air temperature Ta > 0°C), thus favouring ice run and ice thrusting onto the shore. Wind blowing from the west causes ice piling. For instance, in late January 2004, on the sheltered basins of the nearby Rügen Island, a wind blowing from the SW caused the formation of ice piles up to 2 m high along the eastern shore (Schmelzer et al. 2004). On the southern Baltic coastal lagoons, wind blowing from the W sector usually brings warming and causes ice run (cf. Girjatowicz 1977). It causes ice drift and thrusting of sheets of ice onto the shore. Stronger wind may thrust thicker ice sheets that will form higher hummocks (cf. Orviku 1965, Girjatowicz 2014).

Ice scars positioned especially high on trees arise where ice hummocks develop. The thicker the ice floe, the higher wind speed is required to cause ice thrusting and piling. On the other hand, the ice hummock height increases with increasing ice floe thickness. For instance, on the southern Baltic coastal lagoons, 10 cm-thick ice floes may form hummocks up to 3 m high, and 20 cm-thick ice floes may form hummocks as high as 5 m (Girjatowicz 2014).

Ice-induced tree damage was especially abundant on the Rów Peninsula. It is a flat, low-lying area, frequently flooded during high-water events. During periods of strong wind blowing from western directions (SW, W and NW) in conjunction with elevated water levels, an ice field is readily thrust inland, thus causing damage to tree trunks. Figure 10 shows two separate ice thrusts onto the shore of the Rów Peninsula. Map A shows the range of the thrust and ice hummocking caused by the wind blowing from the WNW. Map B shows the same for the wind blowing from the SW. In this part of the lagoon, especially in the vicinity of Skoszewo and Zagórze villages, in recent years, the ongoing shore erosion caused the destruction of the reed belt and subsequently – also of the tree belt, up to several meters wide, including the ice-scarred trees.

Fig. 10.

Relatively large damage was caused on the northern shore. In this segment, ice scars on trees occurred especially high (Table 2). This shore segment is exposed to frequent wind blowing from the S–SW and W directions, which causes the breakup of fast ice. For this segment of the shore, this is a landward wind, causing ice thrusting onto the shore. Additionally, wind blowing from this direction causes warming (Koźmiński et al. 2004), which in turn causes ice run (Ta > 0°C). Damaged trees were located at a relatively short distance from the water line due to the dominance of steep shores in the northern part of the lagoon, which does not favour ice penetration inland. The NW segment of the shore, in the vicinity of Paprotno (a district of Świnoujście town), is exposed almost perpendicularly to the wind blowing from the SW direction. As the most frequent wind direction is from the SW, such shore exposure favours ice thrusting onto the shore.

Numerous ice scars characterised by moderate elevation and located on trees growing relatively far from the waterline were observed on the southern shore. These were formed during periods of strong wind blowing from the W through NW to N, causing the largest water surges on the lagoon, reaching 0.7–1 m, and storm waves (Kowalewska-Kalkowska, Kowalewski 2005, 2008). For comparison, on Lake Ontario, wind blowing at 8 m·s^-1 caused the formation of ice piles reaching a height of up to 2.5 m a.w.l. on the shore (Gilbert, Glew 1986).

The fewest (only five) tree scars formed due to ice activity were identified in the western segment of the shore (Fig. 9). Ice scars on trees were located at exceptionally low elevations (Table 2) and were the smallest (Fig. 8). Sparse occurrence and a short distance of ice scars from the water line may be explained by the basin physiography and the associated wind conditions. The western part of the Szczecin Lagoon is characterised by strong land exposure, a relatively small area and a short fetch of wind from the western sector. The exposure of this part of the lagoon to wind from the eastern sector does not favour ice thrusting on land. Such wind brings cooling (Ta < 0°C) and stabilization or increase of thickness of an ice cover. Such ice cover is stable and lasts for a long period. In spring, it rots and disintegrates in situ (in water).

Numerous papers describe the results of sheet ice piling and thrusting onto lake shores and river banks and their impact on the nearby vegetation, including trees (e.g. Barnes et al. 1994, Engström et al. 2011, Lind et al. 2014). Maximum ranges of ice on shore may be analysed based on the identification of ice scars left behind on trees. Few researchers have applied this method to date (Lederer, Garver 1996, 2000, Pawłowski 2005, Lemay, Bégin 2012, Vandermause et al. 2021). Damaged vegetation and trees growing significantly above the river level are the most frequent evidence for the elevations reached by ice piles on rivers (Lederer, Garver 1996, Uunila, Church 2014, Vandermause et al. 2021). Ice scars on trees have been used to interpret the maximum time of breaking and piling of the ice cover on the rivers (Smith, Reynolds 1983, Vandermause et al. 2021).

The studies by Leder and Garver (1996, 2000), performed along the lower reaches of the Mohawk River (New York), indicated that the ordinate of the ice scar position on trees is equal to or lower than the ordinate of surge culmination during an ice jam. Similarly, on the lower course of the Wisla River, the ordinate of the water level reached during an ice jam is usually matched by the ordinate of the maximum elevation of ice scars (Pawłowski 2005). Ice damage to trees along the banks of the Peace River in Canada was observed up to 2 m above the maximum river level (Uunila, Church 2014). In Alaska, during an ice jam and an associated surge, damaged roots and trunks, as a result of substratum erosion, occurred to an elevation of 1.8 m (Vandermause et al. 2021). On the southern Baltic lagoons, ice hummocks may reach a height of up to 10 m (Zimdars 1941, Girjatowicz 2004) and in sheltered bays of the Baltic Sea, up to 16 m (Slaucitajs 1929, Kraus 1930, Orviku 1965). Along the southern Baltic coast, the dominant form of piled ice are grease ice ridges, whose height may reach up to 5 m (Girjatowicz 2001a, Girjatowicz, Łabuz 2020).

On rivers, water current is the factor responsible for piling ice floes and brash ice to form ice jams (Beltaos et al. 2006, Lind et al. 2014). On coastal lagoons, however, the factor responsible for piling sheets of ice within the ice field to form hummocks is the wind (Girjatowicz 2001b, Orviku et al. 2011, Leppäranta 2013). The influence of the wind (tangential stress) on a large area of an ice field causes the ice to become strongly pressured, to thrust inland and to undergo hummocking (Fig. 11A). The larger the ice field surface and the stronger the wind, the thicker ice will undergo hummocking. Further, higher ice thickness will form higher hummocks (cf. Kraus 1930, Alestalo, Häikiö 1976, Girjatowicz 2014). High hummocks arise due to a specific ice hummocking process. Sheets of ice advancing from the windward side become rafted and may even attain a vertical position. On the leeward side of a hummock, sheets of ice crumble under their own weight, forming brash ice (Fig. 11B). Brash ice accumulating on the leeward side causes a gradual deflection of the ice sheet as it advances through the apical part of the hummock, and thus, the height of the hummock is simultaneously rising. In the final phase of hummocking, the advancing sheet of ice attains a vertical position and may even deflect towards the windward side and crumble into brash ice there as well (Girjatowicz 2019).

Fig. 11.

The highest values of the ice damage ordinate on trees occur when the peak of a hummock aligns with the position of a tree trunk (Fig. 11B). This is where an advancing sheet of ice may damage the tree, marking the maximum elevation of a scar, and at the same time, it is the maximum height of an ice hummock. Ice being hummocked next to such a tree may damage it from the base of the trunk all the way up to the maximum height of the hummock. This is why an ice scar may be so long and reach so high. A tree damaged up to an elevation of 5.5 m by ice thrusting and hummocking due to strong wind over the Vistula Lagoon (Fig. 11B) is a good example of this. This scar marked the maximum height of a hummock. It should be emphasised that neither the scar elevation nor the hummock height should be linked with the water level ordinate during a surge on coastal lagoons. Surges do facilitate sheet ice thrusting onto the shore, however.

The spatial diversity of ice scar occurrence on trees is affected also by shore morphology, in addition to wind direction (cf. Pawłowski 2005). The shores of the southern Baltic coastal lagoons are mostly flat, with extensive areas of swamps, peatlands or marshes. On such shores, sheets of ice may advance over distances of even 100 m inland (Alestalo, Häikiö 1976, Leppäranta 2011, Girjatowicz 2015). Thus, ice scars may occur relatively far from the shore and, usually, at the base of a tree trunk.

The lagoon water level also has a major influence on the distance over which sheets of ice may be thrust inland (cf. Alestalo, Häikiö 1976, Orviku et al. 2011, Girjatowicz 2015). At low and moderate water levels, the advancing ice is usually hummocked at the lagoon shore. Under such hydrological conditions, trees are usually damaged only very close to the shoreline. At higher water levels, however, trees may get damaged further away from the shore. In such cases, due to reduced friction, ice thrusting is facilitated, and sheets of ice may advance further inland.

The ongoing climate change, manifested for instance by increased wind velocity and frequent alternations of positive and negative air temperatures, influences the high variability and dynamics of ice phenomena (cf. Jevrejeva et al. 2004, Livingstone et al. 2009, Sztobryn et al. 2012, Haapala et al. 2015, Girjatowicz, Świątek 2021). A given winter may witness several periods of ice phenomena occurrence on the lagoons, for instance, six on the Vistula Lagoon (Lazarenko, Majewski 1975) and up to eight uninterrupted series of days with ice on the Szczecin Lagoon (Girjatowicz, Świątek 2020). Repeated ice cover disintegration, caused usually by strong wind, enables repeated ice damage to trees in individual winter seasons.

Based on analysis of the occurrence of ice scars on tree trunks on the Szczecin Lagoon shores performed here, the following conclusions are drawn: –

ice scars on trees may be a proxy for the dynamics of ice phenomena, especially concerning the distance over which ice is thrust inland, and the height of ice piles on shores,