*Article* **The Evolution of Gravel-Bed Rivers during the Post-Regulation Period in the Polish Carpathians**

**El ˙zbieta Gorczyca 1,\*, Kazimierz Krzemie ´n <sup>1</sup> and Krzysztof Jarzyna <sup>2</sup>**


Received: 13 December 2019; Accepted: 14 January 2020; Published: 16 January 2020

**Abstract:** This study provides a conceptual model of the functioning of gravel-bed rivers during the post-regulation period in Poland and forecasts their subsequent evolution. The main difference between fluvial processes during the pre-regulation and post-regulation period is that they are limited to a zone that is currently several times narrower and trapped in a deep-cut channel. During the river post-regulation period, the construction of additional river training works was significantly limited in river channels. Moreover, all forms of economic activity were significantly reduced in the channel free migration zone, particularly bed gravel extraction operations. As a result of these changes, a limited recovery of the functioning and hydromorphology of the river channel occurred via a return to conditions in effect prior to river regulation. In recovering sections of river, the channel gradually broadens, and its sinuosity and number of threads increase. The overall process can be called spontaneous renaturalization, which yields a characteristic post-regulation river channel. The conceptual model was developed on the basis of the evolution of the gravel-bed river, the Raba River, during the post-regulation period in the Polish Carpathian Mountains.

**Keywords:** river channel; spontaneous renaturalization; post-regulation period; river training; Polish Carpathians

#### **1. Introduction**

Since the late 19th century, gravel-bed river channels in mountain areas and their forelands have been subject to strong human impact in the temperate climate zone. This is due to construction of levees, river training works, channel debris extraction, and construction of reservoir dams [1–3]. River channels have in effect changed substantially in terms of cross-section and longitudinal profile.

Similar change processes have been observed in the Polish Carpathians and their foreland areas [4,5]. Direct human impact on river channels has been accompanied by changes in land use, which varied from catchment to catchment. Decades of intensive river regulation have led to a simplified morphology and pattern of Carpathian rivers which, consequently, became more narrow and deeper, in addition to following a single thread instead of many, with little ability for the river channel to migrate [4,6–8]. Increasing channel depth due to human impact [9–11] has led to large changes in connectivity between the river channel and the floodplain. The outcome in terms of ecology has been the loss of the most valuable habitats along the channel including initial gravel bar ecosystems and alluvial forests in mountain areas [12]. The regulation of the Carpathian river channels has also had other negative outcomes including a reduction in their ability to retain water that could be used for commercial or farming purposes, as well as an increase in the erosion risk to bridges, buildings, and other structures found near the river channel [13].

The period of the largest changes in channel morphology and overall functioning of Carpathian river channels caused by river training works and channel debris extraction lasted from 1961 to 1990. Today the drive to change river channels is much less substantial in the studied area. In addition, it is illegal to collect the gravel from river channels and various government agencies have allowed the development of the free migration zone. The period after 1990 is called herein the post-regulation period. Changes in river channels observed during this period of time have been termed spontaneous renaturalization [14]. Such changes occur most often in the Carpathian river channels with a high width to depth ratio (W/D) [8]. Similar evolving patterns are present in the channels of other rivers, for example, the Sacramento River in California, USA [15] and the Magra and Piava rivers in northern Italy [16,17].

Researchers perceive river channel renaturalization either as a hydro-geomorphologic process [5], biological process [18], or ecological process [19,20]. The authors of this paper treat the examined subject matter from a hydro-geomorphologic perspective.

The main purpose of this paper is to discuss the pattern of spontaneous renaturalization in a mountain area gravel-bed river channel. This process is believed to represent an important mechanism that explains the functioning of a channel system during the post-regulation period in Poland. The course of renaturalization is described for the example of the Raba River flowing across southern Poland. The river serves as a good representation of Carpathian gravel-bed rivers. River renaturalization is also discussed in terms of future change directions and rates of change.

The Introduction of this paper focuses on the chronology of river channel change as well as types of human impact in the Raba river channel in the 19th and 20th centuries. This history of human impact led to the creation of a regulated river channel by the 1970s. The paper focuses on the analysis of channel morphology and functioning during the post-regulation period using the 1970s and 1980s as a base period for the purpose of comparison. The changes discussed in this paper occur as a result of fluvial processes that vary in intensity from site to site depending on the local fluvial environment, including type and degree of human impact in the channel. Special attention is given to the role of natural fluvial processes that, in particular, include large floods which are considered to be the most important part of successful spontaneous renaturalization processes in the Raba River.

Large floods occurred along the Raba River every two to four years in the period from 1951 to 1972. These major flooding events affected the morphology and the channel pattern of the Raba River including the years 1951, 1958, 1960, 1962, 1963, 1965, 1970, and 1972. The period from 1973 to 1996 was characterized by a much smaller frequency of large and very large flood events. The only major floods in this period to affect the morphology of the Raba River occurred in 1987 and 1996. The year 1997 initiated a new period of increasingly frequent floods, of which the most critical events in terms of channel morphology change were the floods of 1997, 2001, 2010, and 2014 [8].

The analysis of flood effects on changes of functioning in the Raba river channel in post-regulation period employed the concept of the flood pulse [21,22]. This concept does not focus on the river channel but its relationship with its overall environment including the valley floor [23]. The concept assumes that a river channel follows an oscillating pattern of widening and narrowing. The former occurs most intensively after large flood events, which revitalize lost connectivity between the river channel and its floodplain and allow for a regeneration of hydro-geomorphologic processes [24–26]. Looking for proof of the formation of channel widening zones following floods and their narrowing during floodless periods constitutes an important part of this study.

The analysis of morphologic diversity in the longitudinal profile of a post-regulation channel in this paper uses the concept of beads on a string. This concept was proposed by Stanford et al. [27] and Ward et al. [28] and introduces the idea of beads representing zones of the most intensive morphologic, hydrologic, and biological diversity based on the capacity for the formation of a broad floodplain. These zones are characterized by multidimensional interrelationship between the river channel and the floodplain, with one affecting the other. Bead zones alternate with string zones that have a simplified structure and limited ability to affect their surroundings. In this study we identify bead zones along the Raba river channel, analyze conditions for their occurrence, and assess their significance in spontaneous

renaturalization. This is another important part of the proposed conceptual model of the evolution of a gravel-bed river in a mountain area during a post-regulation period.

#### **2. Study Area**

The Raba River is one of the larger mountain area rivers in the upper Vistula basin. It is 131.7 km long, with an average gradient of 0.25◦. The area of the Raba catchment is 1537 km2. It is an elongated catchment, with a larger proportion of right-bank catchment area (Figure 1). The Raba catchment is located in the following two different geomorphologic regions: (1) the Carpathian Mountains and (2) the Carpathian foreland basins (Figure 2). The study area is formed of flysch rocks from the Cretaceous and Paleogene, in its Carpathian part. These are mostly sandstone, claystone, siltstone, and conglomerates [29]. Sub-Carpathian basin, on the other hand, is formed of thick clastic formations from the Miocene that lay atop Mesozoic, Paleozoic, and Precambrian formations to yield an unconformity. The Miocene formation includes mostly siltstone, claystone, and shale with sandstone inserts [29]. The Raba river channel cuts into relatively thin alluvial deposits or flysch parent material [8]. The highest point in the catchment is Mt. Turbacz (1310 m). The lowest point is the river's point of confluence with the Vistula River at 180 m above sea level. The lowest annual precipitation totals are noted in the northern part of the catchment across the Carpathian foreland and range from about 700 to 800 mm/yr. The highest totals are noted in the middle mountain area of the Gorce Range and equal about 1200 mm/yr. The mean annual discharge of the Raba River at the Proszówki gauging site covering more than 90% of the catchment area equals 17.6 m3/s, while the maximum recorded discharge equals 1470 m3/s [30].

**Figure 1.** Study area.

The Raba catchment is a rural area characterized by the predominance of the agricultural type of land use [31]. Arable land covers 40% of the catchment area, forests cover 35.6%, meadows 12%, and fruit growing areas about 2%. Built-up areas cover the remaining 10.4% of the catchment area [32]. Many municipalities are located in the near vicinity of the Raba River and produce a significant impact on land use and management across the valley floor. This consequently affects the functioning of the channel. The high human impact on the valley floor and river channel is due to the presence of vital transportation links such as roads and railways. One of the most important roads in southern Poland runs through the Raba valley.

**Figure 2.** Location and intensities of downcutting in the Raba river channel.

The Raba river channel has been gradually regulated since the beginning of the 20th century. Today about 60% of the river channel is lined with longitudinal structures, which were built during every decade since the 1960s. About 600 m of regulated banks are noted, on average, for each one kilometer of the Raba river channel. About two-thirds of the structures were deemed to be well preserved, while the rest as poorly preserved. River training works are found most often along concave banks, and less often along convex banks. The structures are meant to protect against erosion. In areas characterized by a high density of buildings situated close to the riverbanks, both banks are usually regulated in order to provide flood protection.

Transverse training structures such as straight drop structures are much less impactful than longitudinal structures.

The most important hydrotechnical structure found along the Raba River is the dam at Dobczyce. The resulting Dobczyce Reservoir was built in 1986 and is located at the 60th kilometer of the Raba River. The purpose of the reservoir is to help local communities obtain drinking water, protect local communities from flooding, equilibrate river discharge, and generate electricity.

#### **3. Research Methods**

Research on channel structure of the Raba River was performed by employing a channel mapping procedure developed at the Institute of Geography and Spatial Management, Jagiellonian University [33,34]. The main assumption employed in this research method is an understanding of the river channel system as a structure consisting of homogenous morphodynamic reaches. The log used to collect data covers the following four groups of data: (1) initial information, (2) channel parameters, (3) hydrodynamic parameters, and (4) hydrometeorologic parameters of the studied period of time. Principal log entries included information on geographic location, geologic structure, channel parameters, cross sections, longitudinal profile, banks, floor type, sediments, buildings along the watercourse, and channel type. The data log included 57 types of quantitative information and 48 types of qualitative information. Quantitative information is used to calculate indices characterizing the studied channel [34]. The entire river channel of the Raba River was surveyed in line with the above guidelines, in 2015. In addition, some reaches were surveyed in 2009, 2010, and 2014. Grain-size distribution of bed material were assessed using a method proposed by Wolman [35].

Information on the degree of river regulation works and their current state was also collected during the field surveys and supplemented with corresponding information from the Regional Water Management Board in Kraków.

The Raba river channel was also examined using a digital elevation model produced from 2010 to 2013. Orthophotos were also examined for the years 2009 and 2012. These are available from the Cartography Collection at Institute of Geography and Spatial Management, Jagiellonian University and were received from the Central Geodesic and Cartographic Documentation Bureau. The analysis in the study was enhanced using orthophotos available from Google Earth for the period 2014 to 2019. Cartographic analysis and morphometric analysis were performed using QGIS software.

Data provided by the Institute of Meteorology and Water Management (Polish, IMGW) were used to examine the hydrodynamic characteristics of the studied watercourses. Six gauging sites were analyzed in total (Figure 1).

This study compares survey data for the Raba river channel from the years 1976 to 2015. A survey of the Raba river channel performed in 2015 utilized the boundaries of homogenous morphodynamic reaches identified during the 1976 survey. Therefore, the zones of spontaneous renaturalization ("beads on a string") identified in the Raba river channel in 2015 do not overlap with the boundaries of the analyzed channel reaches. The procedure for the survey performed in 2015 was described earlier in this paper. In 1976 the Raba river channel survey was performed as part of a project called "Dynamic river channel typology for the Carpathians and their foreland." The 1976 project was run by the Institute of Geography of Jagiellonian University, Department of Geomorphology [36]. The survey data on the Raba river channel morphology and pattern from 1976 make it possible to make viable comparisons with analogical survey data collected in 2015. The data logs used in 1976 were somewhat different with respect to selected parameters, i.e., more descriptive and less quantitative. This means that not all information contained in the data logs could be compared. In this study we also looked, in detail, at the morphology and pattern of the Raba river channel, and changes therein, for major flood events in 2010 and 2014. The observed changes make it possible to determine change directions in the evolution of gravel-bed rivers in mountain areas large floods.

#### **4. Human Impact on Carpathian Rivers**

Carpathian river channels that were regulated starting in the late 19th century had evolved in earlier periods to high bedload transport conditions resulting from intensive agriculture and the development of a dense network of unpaved roads. In the 19th century the Carpathians were a densely populated region, largely deforested, with arable land found as high as 1000 m above sea level. The region had a very high road density at the time as well. Today, unpaved road density ranges from 4 to 7 km·km−<sup>2</sup> in the Carpathians [37]. In the late 19th century, unpaved roads were utilized more intensively and were an important source of both fine and coarse material as they were eroded as deep as bedrock [38]. If we were to consider channel evolution for periods prior to the period of high human impact (regulation works and large scale gravel mining), then we could argue that river channels in the Western Carpathians were multithreaded, wandering, or sinuous with lateral bars, and meandering in basins with a tendency for braiding [10,39,40].

Flood control levees have been built in Carpathian river valleys since the late 19th century. As a result, the channel migration zone has become smaller over time for rivers across the region. Major engineering works began in Carpathian river channels in the late 19th century. The aim was to regulate river channels in order to reduce flood risk and create commercial waterways. River regulation work proceeded in stages starting with reaches at lower elevations, and then shifting upriver [41]. Regulated river channels had gentler bends and longer straight sections [42]. River regulation work in the Carpathians can be divided into several distinct stages which are the following: (1) The period before World War II (described above), (2) the period from 1945 to 1989, and (3) the period after 1990. Each stage varied in terms of the method of regulation, extent of regulation, and intensity of regulation [4,8].

Stage 1 consisted of extensive work on long sections of river and some projects were never fully completed. Regulation work at this stage consisted of watercourse straightening, as well as a narrowing and shortening. Some Carpathian rivers were made shorter by as much as 10% [8,43]. Regulation projects were halted by the outbreak of World War II and resumed only in the late 1950s. In the beginning of Stage 2, since the 1940s to the 1960s, industrial-scale extraction of channel debris was introduced in Carpathian rivers [5]. In later decades, the local population continued to illegally extract debris from the river channels in the Carpathians [6,8]. The most advanced regulation works were performed in the 1960s and 1970s in the later part of Stage 2 [4]. This was a period of the most radical alteration of river channel patterns, longitudinal profiles, and cross-sections. This period produced mainly such river training structures such as groins, flow redirection piers, weirs, and debris dams, but few drop structures or steps.

In the 1970s, in Poland, river regulation works were planned at the local level [44]. The number of regulation projects decreased markedly in the 1980s, with drop structures being the main form of river regulation [6]. Some banks were also reinforced in this period of time. The main reason for this drought in investment was a major economic crisis in Poland, as well as the lack of major flood events. Regulation works became limited in scope when Poland made the transition to a market economy in the 1990s. Then, the river regulation Stage 3 began in the Polish Carpathians. New regulation works were initiated in the late 1990s, especially following a major regional flood in 1997. River engineering projects at the time were limited primarily to repairing damage produced by floods in 1997, 1998, and 2001. Work included the building of drop structures and bank reinforcements. Narrowed and straightened river channels lacked debris, which had been previously supplied by unreinforced banks, leading to an acceleration in downcutting erosion that undercut the base of existing hydrotechnical structures. The next stage would include the destruction of these structures. Given such outcomes, river engineers would reinforce the same banks two or three times [8].

River regulation systems built in the Carpathians turned out to be ineffective and harmful to the natural environment. Regulation works aimed to produce a single, narrow channel with no side channels. The large change in river channel geometry led to a lack of adjustment when paired with a river's actual hydrologic regime. Downcutting increased in river channels as a result [6,8,11,45,46]. In some cases, downcutting reached solid rock or damaged river engineering structures [4,11,47]. This forced new river regulation work to be performed. This repetition would eventually lead to extensive degradation in river channels, as much as 80% in the Raba river channel [8].

The gradient of regulated rivers would increase along straightened, shortened, and narrowed reaches, as would discharge and the rate of debris transport, which would lead to a deficit of debris in the river channel and, subsequently, a lowering of the channel floor [1,8]. Channel incision by as much as 1.3 to 3.8 m has occurred in the lower and middle reaches of Carpathian rivers since the early 20th century [9,11]. In the second half of the 20th century, channel floor lowering has been observed in upstream reaches of Carpathian rivers, ranging from 0.5 to 3.5 m [8,13,48,49].

In light of the general change in approach to river channel maintenance in the Carpathians, in Poland after 1989, human impact in river channel systems has been severely limited since. In place of extensive intervention, the strategy now is to pursue a minimal intervention focus, i.e., one that is close to what nature itself would pursue. The stoppage of river regulation work in the Carpathian region has led to spontaneous renaturalization of river channels [8].

### **5. Results and Discussion**
