**1. Introduction**

Flooding is one of the most damaging natural hazards worldwide, with flood risks ever increasing [1]. This can, among other things, be linked to climate change, which is likely to further intensify flooding. Detailed evaluations of climate change impacts on flooding at local level, however, are still inconclusive [2,3]. A second important driver of increasing flood damage is socio-economic development, including an increase in wealth, population development, and settlement expansion in flood-prone areas [4–6].

In order to reduce flood risks, a shift from controlling rivers and hazard prevention to an integrated approach of flood risk managemen<sup>t</sup> can be observed [7–10]. Integrated flood risk managemen<sup>t</sup> includes structural and non-structural measures, moving away from a solely engineering task to the inclusion of different disciplines [11,12]. With this paradigm shift, spatial planning has become a crucial part of flood risk management, and related policies are seen as an essential means to prevent flood damage [11]. Scholarly literature uses the term 'spatial turn' to describe the increasing relevance of (mainly undeveloped) land for flood risk managemen<sup>t</sup> [13,14]. According to van Ruiten and Hartmann [14], three aspects can be regarded as valid indicators for this spatial turn in flood risk management: "[ ... ] *the policy of more space for the river, an integrated approach to the issue, and an approach*

**Citation:** Junger, L.; Hohensinner, S.; Schroll, K.; Wagner, K.; Seher, W. Land Use in Flood-Prone Areas and Its Significance for Flood Risk Management—A Case Study of Alpine Regions in Austria. *Land* **2022**, *11*, 392. https://doi.org/10.3390/ land11030392

Academic Editors: Matej Vojtek, Andrea Petroselli and Raffaele Pelorosso

Received: 17 February 2022 Accepted: 4 March 2022 Published: 7 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

*beyond structural measures*". To operationalize the spatial turn, Löschner [13] goes into more detail, including, among other aspects, "*policy efforts to widen rivers and improve the connectedness between rivers and floodplains".*

Research has been conducted on the implementation of policies addressing 'making space for rivers' [15]. This includes research on the 'Room for River Directive' in the Netherlands [16–18], analysis of the practical implementation of giving space to rivers in Germany [19], and studying the application of the 'Room for River Directive' in Canada [20]. Furthermore, in the UK, there is research on the effectiveness of national planning policy in achieving compromises between space for rivers and space for people. This underlines the "*conflict between land and water*" [21].

This paper builds upon the aspect of 'making space for rivers', which, in practice, means an increasing demand of land both for river restoration and flood risk management. In consequence, this demand of land for rivers can lead to land use conflicts [21]. Although the implementation of related policies has been widely explored, there is research demand on the quality and extent of land use conflicts in flood-prone areas, particularly in areas with limited space for permanent settlement. This research will address this gap by looking at land cover changes and the current land use in flood-prone areas in Austrian Alpine regions. Considering the aspects of transformation of water bodies, settlement development, and agricultural land, this paper will answer following research questions. How did water bodies and settlements in flood-prone areas change between 1826–1857 and 2016 in Austrian Alpine regions? How are flood-prone areas in Alpine regions currently used (focusing on settlements and agriculture)?

A historic view on the human transformation of Alpine water bodies shows the decreasing space for rivers. As part of flood risk management, structural measures have been applied, including river channelization. Research on the channelization of Alpine rivers shows that, in comparison to other European countries, rivers in Austria have been altered more intensely [22]. These regulations left rivers with limited land availability [23]. Thus, river regulations are the preceding development, making it necessary nowadays to consciously implement policies in order to return space to rivers. However, it is not simply narrowing riverbeds that affect runoff. The use of flood-prone areas for settlements has an additional impact on discharge due to an increase in sealed surfaces water that cannot infiltrate into the ground, impacting the flood event [23]. Besides the relevance of widening rivers as a part of flood risk management, the ecological benefits must not be ignored [24]. The use of floodplains for agriculture as well as settlements often has a negative impact on the aquatic ecosystems [25,26].

Furthermore, a view on settlement development in flood-prone areas will underline the increasing land-take and will show the land use pressure in flood-prone areas. Research has been conducted on the exposure of settlements to flooding on different levels, showing a general increase in built-up land in hazard zones [5,27–29]. Our research will add to this broad spectrum of research on flood exposure by analyzing settlement development in areas up to a one in a 300-year flooding on a cross-regional level.

Besides settlement development, this paper presents an analysis of agricultural land in flood-prone areas, with a focus on land valuable for regional food production. A considerable share of areas in flood-prone areas is used for agricultural purposes [30]. The potential for mutual impacts is accordingly great, which makes it important for inclusion in this research.

The paper aims at showing land use dynamics in flood-prone areas and the resulting consequences for flood risk management, including the limitations of giving space to rivers.

#### **2. Materials and Methods**

#### *2.1. Study Area*

The research presented in this paper focuses on Alpine regions in Austria, and even though they are not homogeneous, they share similar spatial challenges. First of all, Alpine areas are prone to multiple natural hazards, including snow avalanches, landslides, and floods; floods, however, cause the most economic damage [31]. Furthermore, the space for permanent settlement is limited due to topography. According to the Alpine Convention, 65% of Austrian territory is part of the Alps. However, only 35.7% of the area for permanent settlement is situated within the Alps. This results in only 21.18% of the Austrian Alps being suitable for permanent settlement [32]. In addition to the limited space, the on-going conversion of (mainly) farmland to housing, commercial and industrial areas, and traffic and recreation infrastructure increases the pressure on open-space land uses. For Austria, this so-called land-take amounted to approximately 12 hectares per day in 2020 [33]. The increasing land-take in combination with limited space for permanent settlement particularly exerts pressure on existing and available land resources for flood runoff and flood retention [34].

The research is based on GIS analysis of spatial data on the land use of flood-prone areas, including settlements, water bodies, and agricultural land. For water bodies and settlements, historical as well as current data exists for the catchment area of the rivers Rhine (Vorarlberg), Salzach (Salzburg), and Drava (East Tyrol and Carinthia) (see Figure 1). The agricultural land use was analyzed based on current data, which was available for the whole of Austria.

**Figure 1.** Study area (source of borders: Federal Office of Metrology and Surveying (BEV), Alpine Convention; source of digital terrain model: GELAENDE—basemap.at, accessed on 28 February 2022).

## *2.2. Data*

The historical settlement data and the data on water bodies are based on the Franziscean Cadastre (1826–1859), which was digitalized by Hohensinner et al. [35]. The use of historical data can lead to some inaccuracies; the Franciscean cadastre shows variations in accuracy between intensely used valley areas and more remote locations higher up [35]. For the analysis of settlements, the lesser details at high altitudes does not matter as much, because the main settlements were located in the valleys.

For current land use (2016) based on LISA (Land Information System Austria) data, several datasets were combined [36]. To add detailed differentiation within agricultural land uses, the IACS (Integrated Administration and Control System) dataset was used [37].

Further, the glaciers are based on Buckel and Otto [38], and a dataset on mid-sized rivers and smaller ponds was gathered from OpenStreetMap [39]. For small rivers and streams, data from a project on "Strategic Planning for Alpine River Ecosystems" were used [40]. The different datasets were combined and prepared by Hohensinner et al. [35]. For the recent agricultural land analysis, the IACS spatial dataset of 2018 was used. For the exposure analysis of particularly valuable agricultural land, the results of the BEAT project were used as a spatial dataset, comprising the areas of valuable agricultural land [41]. For the economic evaluation of agricultural land, regional standard output coefficients were considered [42]. The standard output represents the average monetary value of agricultural output at farm-gate price in euro per hectare or per head of livestock. In order to quantify the decline of agricultural land in recent decades, the data of the agricultural structure survey and IACS data from the Green Report 2021 were used [43].

Flooding area combines information derived from the existing flood hazard maps provided by the Federal Water Engineering Administration (Bundeswasserbauverwaltung) and the Austrian Service for Torrent and Avalanche Control (Wildbach- und Lawinenverbauung), as well as data on 200-year flood events from HORA (Austrian Flood Hazard Map). The flood-prone area used for the exposure analysis encompasses all available flooding data in Austria, showing the area potentially affected by a 300-year flood event, which will be referred to as the flood-prone area in this paper. The focus is put on a 300-year flood, which includes areas with lower probability of flooding, in order to integrate areas that might be affected more by flooding in the future due to climate change impacts. Additionally, by covering an area of a 300-year flood in this research, 'protected' areas behind structural flood protection measures are included in the land use analysis of flood-prone areas. These data were also prepared by Hohensinner et al. [35].

## *2.3. Methods*

#### 2.3.1. Human Transformation of Alpine Water Bodies

In order to ascertain human modifications of the Alpine channel network, the historical and current active channels, i.e., water-covered areas and adjacent unvegetated sediment bars, were vectorized using ESRI ArcGIS 10.6. A comparative analysis yielded quantitative losses of running waters as a consequence of river channelization and flood prevention measures. The areal changes in standing water bodies (lakes, ponds, reservoirs) were derived analogously [35].

#### 2.3.2. Settlement Development

The settlement area for this research includes buildings and adjacent streets, squares, gardens, and smaller parks. The method applied to research the development of settlements inside flood-prone areas is described as a "*classical approach for flood exposure assessment*" by Papilloud et al. [44], which has been broadly applied. Using ESRI ArcGIS 10.6., the settlement development was calculated by overlapping the historical with the current data; the difference between the datasets represents the development. To calculate the exposure of settlements to flooding, the results of the first step (settlement development) were further crossed with the flood-prone area. This made it possible to see the difference between the development inside and outside flooding areas. Besides the temporal aspect of the analysis, it was also interesting to look at the regional differences. For this purpose, the data were intersected with the municipality borders, showing settlement development inside and outside flood-prone areas per municipality. The results from the GIS analysis were further processed in Excel to calculate the relative development, which is based on the factor calculated by dividing the current settlement area by the historical area.

The historical data in East Tyrol and the eastern part of the province Carinthia was less detailed than in the other regions, and settlements included grassland to some extent [35]. In order to correct the settlement extents, as an estimate, half of the historical settlement area was excluded from the investigation based on a previous analysis of local sample data. The results of some individual municipalities may therefore be less accurate. Even though

the analysis was broken down to a municipality level, the spatial distribution was observed on a regional level. In addition, due to some spatial inaccuracy between the historical and the current data [35], the analysis worked with the assumption that settlement areas have not been decreasing in a significant manner since the mid-19th century. Therefore, settlement areas that appear in the historical context but do not in the current data were added to the current settlement.

The flood-prone area is based on current calculations but was also used for the historical analysis. It is likely that a 300-year flood in the historical context covered more ground. The results therefore only show the minimum of settlements exposed to a 300-year flood.

#### 2.3.3. Agricultural Land

In the case of agricultural land, the analysis is based on the IACS dataset of 2018. It allows a detailed differentiation of agricultural land use on field level (18 land use categories). Spatially overlapping the land use data with the flood-prone areas shows the agricultural areas affected by flooding. To calculate the economic value of crop production in flood-prone areas, the spatial land use areas in hectares were multiplied by the regional standard output coefficients (according to crop type). This reveals the economic value of crop production within flood-prone areas (livestock production was not taken into account). It allows regionally differentiated conclusions on how strong the pressure might be for farm enterprises and where the situation might be more conflict-ridden between farming and flood risk management. Results of the research project BEAT categorized agricultural land within agricultural production zones based on their relevance for regional food supply [41]. Intersecting flood-prone areas with these data shows potential pressure in the view of food security. The analysis was carried out for Austria and specifically for the Federal State of Carinthia as an Alpine region.
