Evidence for Repeated Re-Activation of Old Landslides under Forest. Van Den Eeckhaut, M., Muys, B., Van Loy, K., Poesen, J., & Beeckman, H. Earth Surface Processes and Landforms, 34(3):352–365, March, 2009.
doi  abstract   bibtex   
Local reactivations of landslides in forests are rarely reported in landslide catalogues. The occurrence of hillslope sections with fresh morphological landslide features in forested old, deep-seated landslides, however, suggests that landslide reactivations are not restricted to residential areas. In this study, a dendrogeomorphological analysis of beech stands was used to investigate the periods of reactivation of a deep-seated rotational slide in the Koppenberg forest (Flemish Ardennes, Belgium). The relation to rainfall and the correspondence to landslide reactivations reported in a nearby built-up area were also analysed. [\n] A dendrometrical study preceding the dendrochronological analysis proved that, compared with the nearby reference site, trees on the Koppenberg forest landslide site were significantly more inclined and showed more knees, indicating that the landslide site has not stabilized yet. As the sampled trees are younger than the landslide, dendrochronology did not allow determination of the year in which the landslide was initiated, but analysis of two different tree ring width parameters (i.e. ring eccentricity and growth change) calculated for trees sampled on the Koppenberg landslide and the reference site proved to be of great help in determining the temporal sequence of landslide reactivation. During the past 80 years, several periods indicative of local reactivations (i.e. 1943-1945, 1949-1952, 1967-1970, 1972-1977, 1979-1981, 1988-1997) were found within the investigated landslide, but delineation of the spatial extent of the reactivations during these indicative periods was not straightforward. These periods generally correspond to years with above-average rainfall. Finally, the fact that at least 34\,% of the years indicative of reactivation of the Koppenberg forest landslide correspond to a year in which a landslide reactivation was reported in the Flemish Ardennes suggests that in built-up areas, apart from anthropogenic interventions, natural triggering factors remain very important. [Excerpt: Discussion and Conclusion] From the results presented in this study, it can be concluded that dendrochronology allows determination of periods indicative of local reactivation of deep-seated landslides. This conclusion was drawn after studying beech stands on a seemingly dormant landslide in the Flemish Ardennes. Prior to a detailed dendrochronological analysis, dendrometry helped in selecting a landslide and reference site, and trees in these sites, appropriate for sampling cross-sections or cores. Although for asymmetric crowns, spiral grain and scars no significant differences were found between the Koppenberg landslide and reference site, the tree phenomena seem more abundant on the landslide affected site. The number of inclined trees, the inclination direction and the presence of knees, however, provide most information (i.e. significant difference between landslide and reference site) on local hillslope stability. Similar results were also obtained in a previous study by Fantucci and Sorriso-Valvo (1999), who found that 78\,% of the trees investigated on landslide sites in Calabria were tilted, whereas no tilted trees or distorted stems were observed in the reference sites. The tilted trees investigated by Fantucci and Sorriso-Valvo (1999) were generally oriented in the downslope direction, indicating that slope movement was mainly translational. In this study, however, a bimodal distribution of inclination direction was obtained, which is typical for a rotational slide that is not completely stabilized yet. This indication of rotational sliding strengthens our field observations, and corroborates conclusions drawn from a detailed photogrammetric analysis (Dewitte et al., in press). [\n] Dendrochronological analysis of trees sampled in the Koppenberg forest showed that determination of phases with local reactivation within old, deep-seated landslides is difficult. Due to the absence of scars caused by landslide debris, of clear reaction wood resulting from severe and sudden tilting and of sprouting of adventitious roots in response to partial burial of stems by debris, the dendrochronological analysis was limited to analysis of variations in ring widths. According to Lang et al. (1999), indicative years obtained with this method should be interpreted with caution. In order to distinguish years with landslide activity from stable years, two different parameters were calculated from the measured tree ring widths (i.e. ring eccentricity and growth changes). The combination of the ring width parameters enabled us to detect periods with a high probability of local landslide reactivation (i.e. 1943-1945, 1949-1952, 1967-1970, 1972-1977, 1979-1981, 1988-1997; Figure 9), but the agreement between the results obtained from eccentricity and abrupt growth reduction was rather low. This may be partly due to the fact that the eccentricity was only calculated for four cut trees, whereas growth changes were also obtained for the cored trees. Moreover, delineation of the spatial extent of reactivations in the different periods was difficult. These periods occurred throughout the lifespan of the beeches, and during or prior to years with above average annual or seasonal rainfall depth. Comparison with results obtained from the reference site showed that distinguishing ring width anomalies related to slope movement from anomalies related to climate, endogenous disturbances such as light competition and genetic quality, or exogenous disturbances such as local hillslope gradient, is not straightforward. Moreover, for wet years, rainfall-related ring width anomalies in tree rings analysed for the reference site do not exclude the possibility that the anomaly observed for the landslide sites was partly related to landslide reactivation. However, some of the potential errors reported by Shroder (1978) are not likely to be present in this study. From our knowledge of the study area, tree ring anomalies caused by other geomorphic processes than landsliding can be excluded, and also external biological and environmental factors such as insects, diseases or fire were not reported by the forest management. [\n] Comparison of the results obtained with dendrochronology with a historical catalogue showed that for the period 1910- 1999 approximately 34\,% of the years classified as indicative of landslide reactivation correspond to a year with a recorded landslide reactivation in nearby built-up area. Hence, apart from human interventions, natural triggering factors remain important on these sites. Moreover, the true percentage is probably higher than the obtained 34\,%, because due to the incompleteness of the catalogue low correlation was found for years prior to 1970. Therefore, similar analyses in areas with more complete historical landslide databases are required to strengthen our findings. [\n] Although landslides under forest are often assumed to be dormant, this study proved that local reactivations within the affected area are likely to occur. With regard to landslide risk prevention, it is important that these landslide sites remain forested, because deforestation and other human interventions will certainly increase the frequency and magnitude of the reactivations. [\n] Finally, this study demonstrates the potential of dendromorphology and dendrochronology to investigate the extent, frequency and type of landslides over the past 80 years in areas for which no other detailed data are usually available.
@article{vandeneeckhautEvidenceRepeatedReactivation2009,
  title = {Evidence for Repeated Re-Activation of Old Landslides under Forest},
  author = {Van Den Eeckhaut, M. and Muys, B. and Van Loy, K. and Poesen, Jean and Beeckman, H.},
  year = {2009},
  month = mar,
  volume = {34},
  pages = {352--365},
  issn = {0197-9337},
  doi = {10.1002/esp.1727},
  abstract = {Local reactivations of landslides in forests are rarely reported in landslide catalogues. The occurrence of hillslope sections with fresh morphological landslide features in forested old, deep-seated landslides, however, suggests that landslide reactivations are not restricted to residential areas. In this study, a dendrogeomorphological analysis of beech stands was used to investigate the periods of reactivation of a deep-seated rotational slide in the Koppenberg forest (Flemish Ardennes, Belgium). The relation to rainfall and the correspondence to landslide reactivations reported in a nearby built-up area were also analysed.

[\textbackslash n] A dendrometrical study preceding the dendrochronological analysis proved that, compared with the nearby reference site, trees on the Koppenberg forest landslide site were significantly more inclined and showed more knees, indicating that the landslide site has not stabilized yet. As the sampled trees are younger than the landslide, dendrochronology did not allow determination of the year in which the landslide was initiated, but analysis of two different tree ring width parameters (i.e. ring eccentricity and growth change) calculated for trees sampled on the Koppenberg landslide and the reference site proved to be of great help in determining the temporal sequence of landslide reactivation. During the past 80 years, several periods indicative of local reactivations (i.e. 1943-1945, 1949-1952, 1967-1970, 1972-1977, 1979-1981, 1988-1997) were found within the investigated landslide, but delineation of the spatial extent of the reactivations during these indicative periods was not straightforward. These periods generally correspond to years with above-average rainfall. Finally, the fact that at least 34\,\% of the years indicative of reactivation of the Koppenberg forest landslide correspond to a year in which a landslide reactivation was reported in the Flemish Ardennes suggests that in built-up areas, apart from anthropogenic interventions, natural triggering factors remain very important.

[Excerpt: Discussion and Conclusion] From the results presented in this study, it can be concluded that dendrochronology allows determination of periods indicative of local reactivation of deep-seated landslides. This conclusion was drawn after studying beech stands on a seemingly dormant landslide in the Flemish Ardennes. Prior to a detailed dendrochronological analysis, dendrometry helped in selecting a landslide and reference site, and trees in these sites, appropriate for sampling cross-sections or cores. Although for asymmetric crowns, spiral grain and scars no significant differences were found between the Koppenberg landslide and reference site, the tree phenomena seem more abundant on the landslide affected site. The number of inclined trees, the inclination direction and the presence of knees, however, provide most information (i.e. significant difference between landslide and reference site) on local hillslope stability. Similar results were also obtained in a previous study by Fantucci and Sorriso-Valvo (1999), who found that 78\,\% of the trees investigated on landslide sites in Calabria were tilted, whereas no tilted trees or distorted stems were observed in the reference sites. The tilted trees investigated by Fantucci and Sorriso-Valvo (1999) were generally oriented in the downslope direction, indicating that slope movement was mainly translational. In this study, however, a bimodal distribution of inclination direction was obtained, which is typical for a rotational slide that is not completely stabilized yet. This indication of rotational sliding strengthens our field observations, and corroborates conclusions drawn from a detailed photogrammetric analysis (Dewitte et al., in press).

[\textbackslash n] Dendrochronological analysis of trees sampled in the Koppenberg forest showed that determination of phases with local reactivation within old, deep-seated landslides is difficult. Due to the absence of scars caused by landslide debris, of clear reaction wood resulting from severe and sudden tilting and of sprouting of adventitious roots in response to partial burial of stems by debris, the dendrochronological analysis was limited to analysis of variations in ring widths. According to Lang et al. (1999), indicative years obtained with this method should be interpreted with caution. In order to distinguish years with landslide activity from stable years, two different parameters were calculated from the measured tree ring widths (i.e. ring eccentricity and growth changes). The combination of the ring width parameters enabled us to detect periods with a high probability of local landslide reactivation (i.e. 1943-1945, 1949-1952, 1967-1970, 1972-1977, 1979-1981, 1988-1997; Figure 9), but the agreement between the results obtained from eccentricity and abrupt growth reduction was rather low. This may be partly due to the fact that the eccentricity was only calculated for four cut trees, whereas growth changes were also obtained for the cored trees. Moreover, delineation of the spatial extent of reactivations in the different periods was difficult. These periods occurred throughout the lifespan of the beeches, and during or prior to years with above average annual or seasonal rainfall depth. Comparison with results obtained from the reference site showed that distinguishing ring width anomalies related to slope movement from anomalies related to climate, endogenous disturbances such as light competition and genetic quality, or exogenous disturbances such as local hillslope gradient, is not straightforward. Moreover, for wet years, rainfall-related ring width anomalies in tree rings analysed for the reference site do not exclude the possibility that the anomaly observed for the landslide sites was partly related to landslide reactivation. However, some of the potential errors reported by Shroder (1978) are not likely to be present in this study. From our knowledge of the study area, tree ring anomalies caused by other geomorphic processes than landsliding can be excluded, and also external biological and environmental factors such as insects, diseases or fire were not reported by the forest management.

[\textbackslash n] Comparison of the results obtained with dendrochronology with a historical catalogue showed that for the period 1910- 1999 approximately 34\,\% of the years classified as indicative of landslide reactivation correspond to a year with a recorded landslide reactivation in nearby built-up area. Hence, apart from human interventions, natural triggering factors remain important on these sites. Moreover, the true percentage is probably higher than the obtained 34\,\%, because due to the incompleteness of the catalogue low correlation was found for years prior to 1970. Therefore, similar analyses in areas with more complete historical landslide databases are required to strengthen our findings.

[\textbackslash n] Although landslides under forest are often assumed to be dormant, this study proved that local reactivations within the affected area are likely to occur. With regard to landslide risk prevention, it is important that these landslide sites remain forested, because deforestation and other human interventions will certainly increase the frequency and magnitude of the reactivations.

[\textbackslash n] Finally, this study demonstrates the potential of dendromorphology and dendrochronology to investigate the extent, frequency and type of landslides over the past 80 years in areas for which no other detailed data are usually available.},
  journal = {Earth Surface Processes and Landforms},
  keywords = {*imported-from-citeulike-INRMM,~INRMM-MiD:c-13709881,~to-add-doi-URL,belgium,deforestation,ecosystem-services,fagus-sylvatica,forest-resources,landslides,precipitation},
  lccn = {INRMM-MiD:c-13709881},
  number = {3}
}
Downloads: 0
About
Documentation
Keywords
Search
Users
Terms and Conditions of Use
Privacy Policy
Sitemap
© BibBase.org 2021