Search

The assessment of uncertainties in landslide susceptibility modelling in a changing environment is an important, yet often neglected, task. In an Austrian case study, we investigated the uncertainty cascade in storylines of landslide susceptibility emerging from climate change and parametric landslide model uncertainty. In June 2009, extreme events of heavy thunderstorms occurred in the Styrian Basin, triggering thousands of landslides. Using a storyline approach, we discovered a generally lower landslide susceptibility for the pre-industrial climate, while for the future climate (2071–2100) a potential increase of 35 % in highly susceptible areas (storyline of much heavier rain) may be compensated for by much drier soils (−45 % areas highly susceptible to landsliding). However, the estimated uncertainties in predictions were generally high. While uncertainties related to within-event internal climate model variability were substantially lower than parametric uncertainties in the landslide susceptibility model (ratio of around 0.25), parametric uncertainties were of the same order as the climate scenario uncertainty for the higher warming levels (+3 and +4 K). We suggest that in future uncertainty assessments, an improved availability of event-based landslide inventories and high-resolution soil and precipitation data will help to reduce parametric uncertainties in landslide susceptibility models used to assess the impacts of climate change on landslide hazard and risk.

Since 2014, a landslide susceptibility model is used by the Geological Survey and Spatial Planning Unit from the Regional Council of Lower Austria to guide decision-making and strategic development in the approx. 19,200 km² province. This existing map (1:25000) has been compiled by using a multi-temporal inventory composed of 12889 slides. In order to obtain the landslide susceptibility model, a generalized additive model (GAM) has been applied, using a large range of predictors. Predictions were performed on the basis of sixteen lithological units. To spatially communicate the landslide propensity, predictions are divided into three categories: low, medium, and high, based on quantiles. By design, the low landslide susceptibility covers 78% of the territory while containing 5% of the landslides. The medium susceptibility class covers 16% of the territory, including 25% of the landslides. The high susceptibility class covers 6% of the territory while containing 70% of the landslides. Although apparently able to correctly predict landslide occurrences over these nearly ten years, this map was never quantitatively evaluated. Since late 2021, a following up review project aims to evaluate how well the existing landslide susceptibility model from 2014 was able to correctly predict the landslides occurring after its implementation. This evaluation is based on landslides that occurred after 2014. Subsequently, the landslide susceptibility will be recalculated, and potential differences between the landslide susceptibility models investigated. To assure fair comparison, an identical methodological design is applied. Changes in the spatial prediction are quantified and explored.Preliminary analysis suggests that the adequacy of the 2014 map to predict future landslides is good but highly determined by the inventories characteristics (i.e., quality and mapping method). For instance, 61% of the landslides coming from a high-quality inventory occur over highly susceptible zones. For a low-quality inventory, this percentage is observed to be rather lower (36%). However, it is also determined that, even for the landslides not occurring in the highly susceptible zone, their locations are rather close to predicted highly unstable zones. For instance, more than 80% of any landslide observations are at least 40m away from a predicted highly unstable zone. The preliminary remodeling of the landslide susceptibility (by including these new landslides) suggests for the regional scale that 88% of the territory remains with the same predicted landslide susceptibility class. However, the arrangement for the individual lithological units might substantially differ. Strategies on how to perform a comparison and updating of landslide susceptibility models are discussed.

In the June of 2009 central Europe witnessed a rampant rainfall spell that spread across populated areas of the country. High-intensity rainfall caused 3000+ landslides in the southeastern Austrian state of Styria, and property damages worth €10 Million. Elsewhere in Austria, flooding amounted to reparations worth €40 Million. Numerous synoptic-scale studies indicated the presence of a cut-off low over central Europe and excessive moisture convergence behind the extreme event. In a warmer climate change scenario, such an extreme precipitation event may manifest into a more intense event due to the higher water holding capacity of air with increased temperatures, but this reasoning may not be so straightforward considering the complex physics of precipitation, more so in a topographically heterogeneous region such as the GAR (Greater Alpine Region).The flooding and landslides caused in the region raise an alarm and thus motivate this study whereby we investigate if the rainfall event did become stronger with time due to climate change compared to how it would have been in a counterfactual (climate change free) past. Here we have deployed the CCLM high-resolution regional model coupled with a statistical landslide model to simulate this event (rainfall and landslides) in a pseudo (surrogate) warming scenario. A marked decrease in rainfall intensity is observed in the simulations for a 1 K cooler climate (pre-industrial past) and the consequent landslide risk could reduce by up to 20%. In the future, depending on the changes in rainfall and soil moisture, the area affected during a 2009-type event could grow by 45% at 4 K global warming, although a slight reduction is also possible.In this novel event-based study, we discuss the results from the lens of attribution perspective - how conditional attribution is much more useful compared to the conventional risk-based approach of attributing extreme events. We develop physical storylines to address the increasing risks of landslides in the region.

With changing environmental conditions, the risk of landslides will also change. For the Styrian basin, Austria, we investigate how storylines of climate and land use/land cover change may affect future landslide susceptibility (2071-2100). Our analysis is based on two extreme rainfall events in Styria in 2009 and 2014, which triggered more than three thousand landslides causing a major threat to the local population and significant damage to settlements and infrastructure.Furthermore, while the number of studies analysing the impact of climate and land use change on landslide dynamics is rising, the assessment of their uncertainties is still often neglected. However, the quantification of uncertainties is not only essential for the development of business strategies and policy interventions, but also for increasing transparency and confidence in scientific analysis. Therefore, we additionally analyse the joint contribution of climate change uncertainty and landslide model uncertainty for the developed storylines of landslide susceptibility.We found for the worst-case storyline (4 K warming scenario) a substantial increase in highly susceptible areas due to much heavier rain. However, the estimated prediction uncertainties were generally high in all storylines. We discovered that the parametric landslide model uncertainty was of the same order as the climate scenario uncertainty, while uncertainties due to internal climate model variability were negligible. With an improved availability of event-based landslide inventories and high-resolution ground data, uncertainties in storylines of landslide susceptibility may be reduced.

Many examples of regional scale statistical landslide susceptibility assessments can be found in scientific literature. A real-life application of these maps for spatial planning decisions is less common. As result of the MoNOE research project (Method development for landslide susceptibility modelling in Lower Austria), a landslide susceptibility map has been created. Since 2014, this map is constantly used by provincial spatial planners and geologists to guide strategic settlement development in Lower Austria (approx. 19200 km²). Resulting from a multi-temporal inventory of 12,889 slides, a generalized additive model (GAM) was applied to predict the landslide susceptibility using a variety of meaningful morphological and geo-environmental predictors. These easily-applicable, local-scale (1:25,000) landslide susceptibility maps consist of three susceptibility classes. The three classes correspond to low landslide susceptibility (covering 78% of all pixels within the study area), moderate (16% of all pixels) and high (6% of all pixels). Although well accepted by the stakeholders, a few important questions recently arise: a) Is this map able to correctly predict new landslide events that occurred after the implementation of this map? b) With the inclusion of these new samples, is the terrain susceptibility still the same? c) If the terrain susceptibility has changed with the inclusion of the unused (partly recently mapped) samples, why and to what extent?By aiming to answer these questions, a review project named MoNEW is currently in place, which has the overall objective to quantify the accuracy of the MoNOE spatial predictions. The new landslides were obtained from two main different sources: 1) recently occurred damage related landslides from a cadaster of landslide events (in German: “Baugrundkataster"), and 2) landslides mapped from hillshades of a high-resolution LiDAR DTM. Based on these new landslides, the final quality of MoNOE will be explored and the landslide susceptibility recalculated to identify potential differences. Therefore, the identical MoNOE methodological design will be applied to ensure comparability and quality control. Changes in the spatial prediction will be quantified and deeply explored.First exploratory analysis has demonstrated that most of the new landslides occurred within the highest landslide susceptibility class, indicating an apparent good ability of the past MoNOE susceptibility model to predict these landslides. Depending on the inventory source, 34 to 64% of the landslides occurred within the higher susceptibility class (this percentage was 70% by design in the original MoNOE project). This variation might be explained by the positional accuracy and mapping methodologies of the new landslides. Additionally, it was observed that most of the new landslides occurring in other less susceptible classes (i.e., “low” and “moderate”) were actually located in close proximity to the highest susceptibility class. Given the applicability scale of the MoNOE landslide susceptibility map (1:25,000), these (mostly very low) quantified distances between the landslide locations and the high susceptibility pixels might be inside of the new landslide mapping accuracy. However, how much the landslide susceptibility of the terrain might change with the addition of these new samples is currently under analysis.

Between 22-26 June 2009, Austria witnessed a rampant rainfall spell that spread across populated areas of the country. High-intensity rainfall caused 3000+ landslides in southeast Styria, and property damages worth €10 Million in Styria itself. Elsewhere in Austria, flooding amounted to reparations worth €40 Million. Numerous synoptic-scale studies indicated the presence of a cut-off low over central Europe and excessive moisture convergence behind the extreme event. In a warmer climate change scenario, such an extreme precipitation event may manifest into a more intense event due to the higher water holding capacity of air with increased temperatures, but this reasoning may not be so straightforward considering the complex physics of precipitation, more so in a topographically heterogeneous region such as the GAR (Greater Alpine Region).The flooding and landslides caused in the region raise an alarm and thus motivate this study whereby we investigate if the rainfall event did become stronger with time due to climate change compared to how it would have been in a counterfactual (climate change free) past. Here we have deployed the CCLM high-resolution regional model coupled with a statistical landslide model to simulate this event (rainfall and landslides) in a pseudo (surrogate) warming scenario. A marked decrease in rainfall intensity is observed in the simulations for 1° cooler climate (pre-industrial past) and the consequent landslide risk is reduced varying across GCMs that were used to derive the boundary conditions from.We discuss the results from the lens of attribution perspective - how conditional attribution is much more useful compared to the conventional risk-based approach of attributing extreme events. The novelty of our approach lies in using a high-resolution convection-permitting regional model for a landslide attribution study.

During 22-24 June 2009, Austria witnessed a rampant rainfall spell that spread across populated areas of the country. High-intensity rainfall caused 3000+ landslides in Feldbach, and property damages worth €10,000,000 in Styria itself. Numerous synoptic-scale studies indicated the presence of a cut-off low over the Adriatic and excessive moisture convergence behind the extreme event. In a warmer climate change scenario, such an extreme precipitation event may become more intense due to higher water holding capacity of air with increased temperatures, but this reasoning may not be so straightforward considering the complex physics of precipitation.Precipitation, as a natural atmospheric phenomenon, is dependent upon the dynamic and thermodynamic characteristics of the atmosphere. While it is safe to say that the thermodynamic characteristics of the atmosphere are relatively easier to simulate with confidence using available global models, the same cannot be said about the dynamics. This can be blamed on the chaotic non-linear behaviour of the atmosphere and problem in resolving sub-grid scale processes that reduce the model accuracy for longer spatial scales.CCLM regional model is used to study this extreme precipitation event. Our setup uses IFS data to calculate initial and boundary conditions for the simulations of the ‘present’ case where our attempt is to recreate the event over the same location as the original event. Further we use CMIP5 global climate models (at the RCP8.5) scenario. In particular, these will be applied in the ‘surrogate climate change’ method. Here, the climate change signals are calculated by computing the difference between the thermodynamic fields of the CMIP5 simulations for the future and the past. These climate change signals are applied to the original fields to obtain the ‘changed’ fields which are used to calculate new initial and boundary conditions resembling a climate-change future. A similar approach is to be applied for the ‘past’ case simulations.The idea behind this experimental setup is to establish a ‘storyline’ for the event as it would have occurred in the past, present and the future. The storyline approach provides an alternative to the traditional probabilistic approach for assessing risk enhancement and can serve to study responses of different mechanisms to climate change. The storyline approach also helps in decision-making as event-oriented risk management is easy for people to perceive and respond to. An associated landslide modelling study, which uses the precipitation output of our simulations as input, looks into the probable increased risks of landslides in the region and will directly aid the lives of those living in Southeast Austria.

Empirical models are frequently applied to produce landslide susceptibility maps for large areas. Subsequent quantitative validation results are routinely used as the primary criteria to infer the validity and applicability of the final maps or to select one of several models. This study hypothesizes that such direct deductions can be misleading. The main objective was to explore discrepancies between the predictive performance of a landslide susceptibility model and the geomorphic plausibility of subsequent landslide susceptibility maps while a particular emphasis was placed on the influence of incomplete landslide inventories on modelling and validation results. The study was conducted within the Flysch Zone of Lower Austria (1,354 km 2 ) which is known to be highly susceptible to landslides of the slide-type movement. Sixteen susceptibility models were generated by applying two statistical classifiers (logistic regression and generalized additive model) and two machine learning techniques (random forest and support vector machine) separately for two landslide inventories of differing completeness and two predictor sets. The results were validated quantitatively by estimating the area under the receiver operating characteristic curve ( AUROC ) with single holdout and spatial cross-validation technique. The heuristic evaluation of the geomorphic plausibility of the final results was supported by findings of an exploratory data analysis, an estimation of odds ratios and an evaluation of the spatial structure of the final maps. The results showed that maps generated by different inventories, classifiers and predictors appeared differently while holdout validation revealed similar high predictive performances. Spatial cross-validation proved useful to expose spatially varying inconsistencies of the modelling results while additionally providing evidence for slightly overfitted machine learning-based models. However, the highest predictive performances were obtained for maps that explicitly expressed geomorphically implausible relationships indicating that the predictive performance of a model might be misleading in the case a predictor systematically relates to a spatially consistent bias of the inventory. Furthermore, we observed that random forest-based maps displayed spatial artifacts. The most plausible susceptibility map of the study area showed smooth prediction surfaces while the underlying model revealed a high predictive capability and was generated with an accurate landslide inventory and predictors that did not directly describe a bias. However, none of the presented models was found to be completely unbiased. This study showed that high predictive performances cannot be equated with a high plausibility and applicability of subsequent landslide susceptibility maps. We suggest that greater emphasis should be placed on identifying confounding factors and biases in landslide inventories. A joint discussion between modelers and decision makers of the spatial pattern of the final susceptibility maps in the field might increase their acceptance and applicability.

Landslide inventories are the most important data source for landslide process, susceptibility, hazard, and risk analyses. The objective of this study was to identify an effective method for mapping a landslide inventory for a large study area (19,186 km2) from Light Detection and Ranging (LiDAR) digital terrain model (DTM) derivatives. This inventory should in particular be optimized for statistical susceptibility modeling of earth and debris slides. We compared the mapping of a representative set of landslide bodies with polygons (earth and debris slides, earth flows, complex landslides, and areas with slides) and a substantially complete set of earth and debris slide main scarps with points by visual interpretation of LiDAR DTM derivatives. The effectiveness of the two mapping methods was estimated by evaluating the requirements on an inventory used for statistical susceptibility modeling and their fulfillment by our mapped inventories. The resulting landslide inventories improved the knowledge on landslide events in the study area and outlined the heterogeneity of the study area with respect to landslide susceptibility. The obtained effectiveness estimate demonstrated that none of our mapped inventories are perfect for statistical landslide susceptibility modeling. However, opposed to mapping polygons, mapping earth and debris slides with a point in the main scarp were most effective for statistical susceptibility modeling within large study areas. Therefore, earth and debris slides were mapped with points in the main scarp in entire Lower Austria. The advantages, drawbacks, and effectiveness of landslide mapping on the basis of LiDAR DTM derivatives compared to other imagery and techniques were discussed.

More than 9000 sinkholes have been documented by the Geological Survey of Thuringia in different lithological units of Thuringia of which many posed a serious threat on life, personal property and infrastructure. While it is clear that they are caused by hollows which formed due to solution processes within the local bedrock material, little is known about the surface processes and dynamics of erosion of the sinkhole visible above ground. The objective of this study was to analyze sinkhole surface dynamics over time with 3D models derived from terrestrial photos by structure from motion and multi-view 3D reconstruction. The sinkhole was surveyed by terrestrial photos on two days with a two months break. During each photo session 84 and 237 photos have been taken from all around the sinkhole. The photos were processed to 3D point clouds using Agisoft PhotoScan and compared using the software CloudCompare and the M3C2 plugin. The resulting point clouds show an area with significant change that covers about 26% of the sinkhole. Toppling and a few erosion processes have successfully been detected with an observed change of up to 10 cm. Nevertheless, for future studies the study design has to be improved regarding the point cloud registration process, a longer observation duration and a quantitative evaluation of the quality of the individual point clouds is pending.

Bell, R., Petschko, H., Rohrs, M. and Dix, A. Assessment of landslide age, landslide persistence and human impact using airborne laser scanning digital terrain models. Geografiska Annaler: Series A, Physical Geography, 94, 135–156. doi:10.1111/j.1468‐0459.2012.00454.xABSTRACTLandslides occur worldwide and contribute significantly to sediment budgets as well as to landform evolution. Furthermore, they pose hazards and risks to people and their goods. To assess the role of landslides, information on their age or persistence (i.e. the length of time the morphological characteristics of a landslide remain recognizable in the terrain) is essential. In this study, the potential of airborne laser scanning digital terrain models (ALS DTMs) is analysed for estimating landslide age, landslide persistence and human impact. Therefore, landslides in two study areas, Swabian Alb in Germany and Lower Austria in Austria, are mapped from hillshades of ALS DTMs and combined with historical information on landslide occurren...

Highlights • We modeled landslide susceptibility with statistical and machine learning techniques. • We evaluate performance, predictor importance, and visual appearance of susceptibility maps. • Differences in model prediction performance were for the majority non-significant. • Consequently, landslide modelers may consider selecting modeling techniques based on additional practical criteria. Statistical and now machine learning prediction methods have been gaining popularity in the field of landslide susceptibility modeling. Particularly, these data driven approaches show promise when tackling the challenge of mapping landslide prone areas for large regions, which may not have sufficient geotechnical data to conduct physically-based methods. Currently, there is no best method for empirical susceptibility modeling. Therefore, this study presents a comparison of traditional statistical and novel machine learning models applied for regional scale landslide susceptibility modeling. These methods were evaluated by spatial k-fold cross-validation estimation of the predictive performance, assessment of variable importance for gaining insights into model behavior and by the appearance of the prediction (i.e. susceptibility) map. The modeling techniques applied were logistic regression (GLM), generalized additive models (GAM), weights of evidence (WOE), the support vector machine (SVM), random forest classification (RF), and bootstrap aggregated classification trees (bundling) with penalized discriminant analysis (BPLDA). These modeling methods were tested for three areas in the province of Lower Austria, Austria. The areas are characterized by different geological and morphological settings. Random forest and bundling classification techniques had the overall best predictive performances. However, the performances of all modeling techniques were for the majority not significantly different from each other; depending on the areas of interest, the overall median estimated area under the receiver operating characteristic curve (AUROC) differences ranged from 2.9 to 8.9 percentage points. The overall median estimated true positive rate (TPR) measured at a 10% false positive rate (FPR) differences ranged from 11 to 15pp. The relative importance of each predictor was generally different between the modeling methods. However, slope angle, surface roughness and plan curvature were consistently highly ranked variables. The prediction methods that create splits in the predictors (RF, BPLDA and WOE) resulted in heterogeneous prediction maps full of spatial artifacts. In contrast, the GAM, GLM and SVM produced smooth prediction surfaces. Overall, it is suggested that the framework of this model evaluation approach can be applied to assist in selection of a suitable landslide susceptibility modeling technique.

Landslide susceptibility maps can be elaborated using a variety of methodological approaches. This study investigates quantitative and qualitative differences between two statistical modelling methods, taking into account the impact of two different response variables (landslide inventories) for the Rhenodanubian Flysch zone of Lower Austria. Quantitative validation of the four generated susceptibility maps is conducted by calculating conventional accuracy statistics for an independent random landslide subsample. Qualitative geomorphic plausibility is estimated by comparing the final susceptibility maps with hillshades of a high resolution Airborne Laser Scan Digital Terrain Model (ALS-DTM). Spatial variations between the final susceptibility maps are displayed by difference maps and their densities. Although statistical quality criterions reveal similar qualities for all maps, difference maps and geomorphic plausibility expose considerable differences between the maps. Given that, this conclusion could only be drawn by evaluating additionally the geomorphic plausibility and difference maps. Therefore, we indicate that conventional statistical quality assessment should be combined with qualitative validation of the maps.

In statistical landslide susceptibility modelling the identification of appropriate explanatory variables describing the predisposing and preparatory factors for the landslides of a given inventory is important. In this context information on the age and the respective land cover at the time of occurrence is beneficiary. The potential of mapping very old (or prehistoric) landslides using LiDAR derivatives has not been analysed yet. Additionally, performing a visual interpretation of derivatives of a single LiDAR DTM it is not possible to assign the accurate age or date of the occurrence of the event to each mapped landslide. Therefore, commonly no information on the land cover at the time of landslide occurrence for these very old landslides (but also for younger ones) is available. The objective of this study is, to estimate the relative age of landslides during the mapping and to explore differences of the recent land cover distribution in the relative ages of the landslides. This is performed to evaluate the sustainability of including recent land cover data into susceptibility modelling. The relative age of the landslides is estimated for each landslide according to its morphological footprint on the LiDAR DTM derivatives and to its appearance on the orthophoto. The different relative ages assigned are “very old”, “old”, “young” and “very young”. The study area is located in three districts of Lower Austria, namely Amstetten, Baden and Waidhofen/Ybbs. The resulting inventory includes 1834 landslides and shows that the “very old” and “old” landslides (60 % of all mapped landslides) are mainly covered by forest (~60 % of all land cover types). We conclude that using this inventory including recent land cover data in the susceptibility model is not appropriate for Lower Austria. There is a potential of mapping “old” or “very old” landslides on the LiDAR derivatives. The absolute age remains unknown.

Landslide susceptibility maps are helpful tools to identify areas potentially prone to future landslide occurrence. As more and more national and provincial authorities demand for these maps to be computed and implemented in spatial planning strategies, several aspects of the quality of the landslide susceptibility model and the resulting classified map are of high interest. In this study of landslides in Lower Austria, we focus on the model form uncertainty to assess the quality of a flexible statistical modelling technique, the generalized additive model (GAM). The study area (15 850 km2) is divided into 16 modelling domains based on lithology classes. A model representing the entire study area is constructed by combining these models. The performances of the models are assessed using repeated k-fold cross-validation with spatial and random subsampling. This reflects the variability of performance estimates arising from sampling variation. Measures of spatial transferability and thematic consistency are applied to empirically assess model quality. We also analyse and visualize the implications of spatially varying prediction uncertainties regarding the susceptibility map classes by taking into account the confidence intervals of model predictions. The 95% confidence limits fall within the same susceptibility class in 85% of the study area. Overall, this study contributes to advancing open communication and assessment of model quality related to statistical landslide susceptibility models.

This study focuses on the comparison of different approaches for landslide susceptibility modelling and is part of the research project “MoNOE” (Method development for landslide susceptibility modelling in Lower Austria). The main objective of the project is to design a method for landslide susceptibility modelling for a large study area. For other objectives of the project we refer to Bell et al. (Proceedings of the 2nd world landslide forum, Rome, 3–7 Oct 2011, this volume). To reach the main objective, the two different statistical models “Weights of Evidence” and “Logistic Regression” are applied and compared. By using nearly the same input data in test areas it is possible to compare the capabilities of both methods. First results of the comparison indicate that in valleys and on south facing slopes the results are quite similar. In contrast, the analysis on north facing slopes shows differences. In the ongoing work the reasons for these differences will be analysed. Furthermore, attention will be paid to finding adequate validation methods for the two modelling approaches.

Landslide inventories, their accuracy and the stored information are of major importance for landslide susceptibility modelling. Working on the scale of a province (Lower Austria with about 10,000 km2) challenges arise due to data availability and its spatial representation. Furthermore, previous studies on existing landslide inventories showed that only few inventories can be used for statistical susceptibility modelling. In this study two landslide inventories and their resulting susceptibility maps are compared: the Building Ground Register (BGR) of the Geological Survey of Lower Austria and an inventory that was mapped on the basis of a high resolution LiDAR DTM. This analysis was performed to estimate minimum requirements on landslide inventories to allow for deriving reliable susceptibility maps while minimizing mapping efforts. Therefore a consistent landslide inventory once from the BGR and once from the mapping was compiled. Furthermore, a logistic regression model was fitted with randomly selected points of each landslide inventory to compare the resulting maps and validation rates. The resulting landslide susceptibility maps show significant differences regarding their visual and statistical quality. We conclude that the application of randomly selected points in the main scarp of the mapped landslides gives satisfactory results.

Landslides threaten most parts of the provincial state of Lower Austria and cause damage to agricultural land, forests, infrastructure, settlements and people. Thus, the project “MoNOE” (Method development for landslide susceptibility modelling in Lower Austria) was initiated by the provincial government to tackle these problems and to reduce further damage by landslides. The main aim is to prepare landslide susceptibility maps for slides and rock falls and to implement these maps into the spatial planning strategies of the provincial state.

Die Nutzung und die Umgestaltung von Landschaften haben sich seit Beginn der Industrialisierung stark verandert. Um die Flachenentwicklung zu erfassen und hinsichtlich nachhaltiger Raumentwicklung regelmasig zu beobachten, braucht es geeignete Mittel. Indikatorensysteme helfen, diesen Anforderungen gerecht zu werden. Es gibt bereits zahlreiche indikatorenbasierte Informations- und Bewertungsinstrumente, die die Flachenentwicklung unter quantitativen wie qualitativen Gesichtspunkten bewerten. Diesbezugliche Untersuchungen decken allerdings aufgrund der Datenbasis haufig nur einen rezenten Zeitraum ab. Die Identifikation von auf langen historischen Zeitraumen basierenden Trends bietet die Moglichkeit, die aktuelle Situation umfassender zu kontextualisieren. In dieser Arbeit analysieren wir die Kulturlandschaftsentwicklung in zwei osterreichischen Gemeinden im Voralpenraumn bzw. im Sudostlichen Alpenvorland uber einen Zeitraum von nahezu 200 Jahren. Dazu wurden aus Katasterkarten, Luftbildern und Orthophotos konsistente, raumlich explizite Landnutzungsdatensatze fur drei Zeitschnitte (1820, 1960, 2015) erstellt. Der Fokus liegt auf der Siedlungs-, Wald- und Landwirtschaftsentwicklung, fur die raumbezogene Indikatoren aus geeigneten Indikatorensystemen zur Beobachtung nachhaltiger Raumentwicklung ausgewahlt wurden. Fur jeden Zeitschnitt wurden die Indikatoren zu einem Gesamtindex des Kulturlandschaftswandels aggregiert und anschliesend verglichen. Das verwendete Indikatorensystem eignet sich, um langfristige Trends anhand von Flachenentwicklungen darzustellen. In beiden untersuchten Gemeinden ist der Gesamtindex heute deutlich geringer als um 1820, was eine negativ gewertete Kulturlandschaftsentwicklung darstellt. Obwohl Unterschiede in den betrachteten Entwicklungspfaden bestehen, weist die Siedlungsentwicklung im Vergleich zur Wald- und zur Landwirtschaftsentwicklung in beiden Gemeinden den starksten negativen Trend auf.