University of Vienna, Austria – T. Glade, C. Promper, H. Petschko CNRS, Strasbourg, France – J.-P. Malet, A. Remaître, A. Puissant CSIC, Zaragoza, Spain – S. Bégueria, G. Sanchez, R. Serrano Final Meeting, 26-27 September 2013, Barcelona : changing pattern of landslide risks as a response to global changes in mountain areas
78
Embed
ChangingRISKS - changing pattern of landslide risks as response to global changes in mountain areas
Presentació per part de Jean-Philippe Malet (CNRS/Institut de Physique du Globe & Laboratoire Image, França) en el marc de l’acte de clausura del projecte europeu CIRCLE 2 MOUNTain co-organitzat per l'Oficina Catalana del Canvi Climàtic durant els dies 26 i 27 de setembre de 2013.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
University of Vienna, Austria – T. Glade, C. Promper, H. Petschko
CNRS, Strasbourg, France – J.-P. Malet, A. Remaître, A. Puissant
CSIC, Zaragoza, Spain – S. Bégueria, G. Sanchez, R. Serrano
Final Meeting, 26-27 September 2013, Barcelona
: changing pattern of landslide
risks as a response to global
changes in mountain areas
KEY RESEARCH QUESTION
Can we identify changes in landslide hazard (susceptibility, frequency,
magnitude) and landslide risks (vulnerability, costs) associated to climate and
landuse change scenarios?
What indicators to express these possible changes?
(modified from Glade & Crozier, 2005)
CONCEPT AND METHODOLOGY
Definition of time series & maps of actual/changing predisposing/triggering factors
Creation of actual and ‘changed’ landslide hazard maps
Definition of ‘changed’ maps of landcover and socio-economic factors
Creation of actual and ‘changed’ landslide risks maps
CONCEPT AND METHODOLOGY
Definition of time series & maps of actual/changing predisposing/triggering factors
Creation of actual and ‘changed’ landslide hazard maps
STAKEHOLDERS
Indicators of changes
Maps
CONCEPT AND METHODOLOGY
Definition of ‘changed’ maps of landcover and socio-economic factors
Creation of actual and ‘changed’ landslide risks maps
Definition of time series & maps of actual/changing predisposing/triggering factors
Creation of actual and ‘changed’ landslide hazard maps
METHOD
Comparative analysis of 2 study areas with different environmental conditions
and different exposures to landslide risks
Stakeholder:
Geological Survey & Spatial Planning and Regional Policy
Division of the Federal Government of Lower Austria
more than 650 landslide events for the period 1950-2012
(using orthophotographs interpretation : 1962, 1979, 1988 and
ALS inventory)
>80% of the events were slides
update for temporal assessment (PhD. C. Promper)
Waidhofen/Ybbs – point data: 1950 - 2012
0
5
10
15
20
25
Lan
dsl
ide
s ac
cord
ing
to a
rea
(%)
square meters
200m² classes
1000m² classes >600m² class
Climate classification: analysis of synoptic weather situations
Landslide triggering factors analysis
RESULTS
(1) polar air-mass penetrating through
the Mediterranean sea (P med)
(2) polar air-mass penetrating through
the Atlantic ocean or the North sea
(Pm)
(3) degraded polar air-mass penetrating
through the Atlantic ocean or the
North sea (Pm d)
(4) tropical air-mass penetrating through
the Mediterranean sea (T med)
(5) tropical air-mass penetrating through
the Atlantic ocean (Tm)
(6) continental tropical air-mass (T cont)
Climate classification: analysis of synoptic weather situations
Landslide triggering factors analysis
RESULTS
When the air mass is crossing the
Mediterranean see (P med or T
med), most of the events corresponds
to debris flows (24 cases out of 32).
This must be put in relation with the
occurrence of strong and fast
thunderstorm in summer
Landslide triggering factors analysis
RESULTS
There are no specific relationship
between the occurrence of slow-
moving landslide and a specific air
mass type.
Nevertheless for a specific type (Pm),
90% of the events are slow-moving
landslides. This can be put in relation
with the seasonal frequency of such
air-masses type in the South French
Alps. Indeed, most of these air
masses are crossing the study area in
Winter or early Spring.
Climate classification: analysis of synoptic weather situations
Landslide triggering factors analysis
RESULTS
Rainfall patterns: antecedent cumulative rainfalls and seasonal patterns
Waidhofen/Ybbs
Barcelonnette
Landslide triggering factors analysis
RESULTS
Rainfall patterns: daily rainfall
•Type A: this situation is characterized by heavy daily
rainfall following a 30-days dry period. For most of the
observed events, this climate situation corresponds to
violent generalized summer storms
•Type B: this situation exhibits the same pattern as Type
A except that the rainfall event has been recorded in a
single rain gauge as the other rain gauges have recorded
nothing or just a small amount of precipitation. In this
case, the convective summer storm is localized in a small
area (typically a single crest)
•Type C: this situation is characterized by heavy
cumulative rainfall distributed over a very rainy period of
30 days. This climate situation characterizes either the
progressive saturation of the topsoil, the rising of a
permanent groundwater table and the build-up of positive
pore pressures.
•Type D: this situation characterizes absence of rainfall
events for a given day of occurrence of a landslides or a
debris-flow event. This can be related to two main
explanations: (a) the triggering is due to a rapid snow
melt without any liquid rain, (b) the rainfall occurs as a
localized hailstorm not recorded by the rainfall station.
Daily rainfall vs mass movement: probabilities of mass m.
occurrence for daily rainfall categories…
Landslide triggering factors analysis
RESULTS
Rainfall patterns: daily rainfall
Calculations based on the antecedent daily rainfall model (Crozier & Eyles 1980; extended by Glade et al., 2000)
The decay of antecedent conditions is based on hydrology data, which represent the drainage of a given catchment
Rainfall thresholds: antecedent rainfall approach
Landslide triggering factors analysis
Certainty
No landslide
Landslide not certain
Landslide certain
Da
ily P
rec
ipit
ati
on
in
mm
[10 Days]
-> Differences between winter and summer months
-> Winter: Less daily precipitation required to trigger
landslides
-> Generally: Precipitation on the day of occurrence is much
more important than the previous antecedent rainfall
-> Landslide triggering rainfall events occurred mostly in
summer when events with relatively high precipitation of
continental depressions occur.
Wallner S. (2012) Niederschlagsschwellenwerte für die Auslösung von Hangrutschungen – in progress
RESULTS
Calculations based on the rainfall intensity threshold method (Caine, 1980; Montgomery et al., 2000)
The rainfall intensity is based on the total amount of rainfall for a given duration (1h; 2h; 6h; 12h; 24h), which may
trigger or reactivate a landslide
Rainfall thresholds: intensity-duration approach
Landslide triggering factors analysis
Peak intensity associated to debris flows and mudslides triggering
-> Events characterized by high rainfall intensity and
short episode duration (i.e. mostly the result of localized
convective storms) will trigger mostly debris flows and
shallow slides in relatively permeable soils (e.g. moraines,
scree slopes or poorly sorted slope deposits).
-> Long rainfall periods characterized by low to moderate
average and peak rainfall intensity (i.e. the result of
multiple and successive storms during a period of several
weeks or months) can trigger or reactivate shallow and
deep-seated mudslides in low permeability soils and rocks
(e.g., black marls, clay-rich material).
RESULTS
Rainfall thresholds: intensity-duration approach
Landslide triggering factors analysis
RESULTS
101
100
101
100
Duration (h)
Intensity (mm/h)
A14d (
mm
)Generalization of I-D model
DAI n ][ 2
1
where:
I, D and as in ID model
An: antecedent n-day precipitation (mm)
1 and 2: constants of the model Cepeda, Nadim, Høeg & Elverhøi (2009)
Rainfall thresholds: intensity-duration approach
Landslide triggering factors analysis
RESULTS
101
100
101
100
Duration (h)
Intensity (mm/h)
A23d (
mm
)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1-0.5
0
0.5
1
1.5
logD
logI
0 = 7.75787+-3.37368 logD+-12.2519 logI
No debris flow
Debris flow
5163.16788.0
232.181 DAI d
275.0297.4 DI
for debris flows, a traditional ID threshold is sufficient
+ triggering rainfall 1 to 9 hrs
+ no need of antecedent rain
for slides, an improved performance is achieved with the IAD model
+ triggering rainfall 3 to 17 hrs
+ need of antecedent rain of 50 days
meteorological paramater maps
landcover maps
landslide inventory maps / rainfall thresholds
conditionning factors maps
1897
2003
quantitative
scenarios ->
WORKPACKAGE 2
Climate:
Use of downscaled times series from
Newtech (Max-Planck-IM), Gach2c
(Meteo-France) & SafeLand (CMCC
S.c.a.r.l) projects 1970-2000 / 2070-2100
Landcover:
DYNA-CLUE model scenarios TE
CH
NIQ
UE
S
Climate modelling
Dataset: station (observed) data
• Network of six meteo stations with
daily precipitation data in the
Barcelonnette area
• Summer Pndq0.9 computed for
each station
• Gridded data, 0.035º resolution
(aprox. 4 km)
• Model CMCC-CM1 (downscaling
driven by ECHAM4-REMO
GCM/RCM)
• Radiative / emmission scenarios:
CMIP5 (control), RCP4.5/RCP8.5
(future)
• Periods: reference (1965–2000)
and future (2001–2050)
Dataset: GCM data
1. Scoccimarro E., S. Gualdi, A. Bellucci, A. Sanna, P.G. Fogli, E. Manzini, M. Vichi, P. Oddo, and A. Navarra, 2011: Effects of Tropical Cyclones on Ocean Heat Transport in a High
Resolution Coupled General Circulation Model. Journal of Climate, 24, 4368-4384.
Building time serie analysis from old maps and cadasters
RESULTS
Building time serie analysis from old maps and cadasters
1830
2004
RESULTS
Building time serie analysis from old maps and cadasters
Barcelonnette
RESULTS
Consequence analysis
RESULTS
The development of exposure maps (a) layers of elements at risk (b) susceptibility map (c)
exposure map.
Consequence analysis
RESULTS
Damages according to land cover in
per cent for Waidhofen/Ybbs
Damaging events according to land cover type related to depth
for Waidhofen/Ybbs
Consequence analysis
RESULTS
Exposure of building area and farms in per cent for the past and future scenarios.
The results of the analysis for the future exposure for building area and farms show a high increase in exposure in the last maps for 2100 for all scenarios. The class of very high exposure remains very low for all scenarios, probably due to certain preconditioning factors e.g. steep slopes. The second highest class remains below 10% for all scenarios. However a clear increase in exposure is marked for the class “low”, summing up to more than 30% in the last time step analysed.
Consequence analysis
Example of result of the ski resort area of Pra-Loup
Puissant, A., Van Den Eeckhaut, M., Malet, J.-P., Hervás, J. (subm). Regional-scale semi-quantitative consequence analysis in the Barcelonnette Region, Southern France. NHESS.
RESULTS
PROGRESS
Consequence analysis
Further developments
- Urban building evolution using Geopensim simulation tool
- Consequence mapping with the modelled landcover scenarios
The proposed indicator has been developed to be enough flexible and generic to be applied to regions with diverse risk exposure and socio-economic specificities, and is designed to be independent of the type of landslide causing the damage.
The method includes the creation of a detailed geospatial database on
attributes of elements at risk, and an evaluation of the model sensitivity to changes in the combination of attributes is proposed. Special attention is given paid to the classification of the potential damage maps. The PDI method allows calculation of an index for physical injury, structural and functional impacts and socio-economic impacts.
RESULTS
RESULTS
Potential Damage Index map produced for the winter season
for the centre of Barcelonnette and the housing of Le Bérard at
Faucon-de-Barcelonnette:
a1. DPI map for Barcelonnette
a2. DPI map for Faucon-de-Barcelonnette
b1. DSF map for Barcelonnette
b2. DSF map for Faucon-de-Barcelonnette
c1. DCE map for Barcelonnette
c2. DCE map for Faucon-de-Barcelonnette
d1. PDI map for Barcelonnette
d2. PDI map for Faucon-de-Barcelonnette.
The indicator maps (i.e. DPI, DSF, DSE and PDI) can be used for purposes such as landuse planning and emergency management decision-making in terms of risk reduction. The method has been elaborated through discussion with various categories of stakeholders (local authorities, risk planners) in the study area, and other stakeholders (rescue teams, individuals and insurance companies) may have interest in consulting and using such type of maps with a straightforward and easily understandable information.
Indicators of changes: landuse/cover changes
Indicators of changes: hydro-climatic
Accurate mass movement rainfall thresholds estimates are one of the most important prerequisite in most landslide hazard assessments. In a general way, definition of hydro-climatic indicators of change is much more difficult to assess than the indicators of changes associated to landuse/cover changes for different reasons. For instance, the climatic settings are much more varying than the landuse/cover characteristics in space and in time. Moreover, these variations are not always well recorded due to a lack of climatic stations at the highest parts/stretches of the study areas.
Another problem is the great heterogeneity of climatic datasets, either for ‘real’ observations (raingauges, radar, etc.) than for Climate Change modelling simulations results (spatial scale varies from 100 to 102 km²). Of course many techniques allow to link and merge these data, but the uncertainties are still too severe. It is still not possible to use them as potential indicators.
RESULTS
Nevertheless, some interesting results have been found out during the project. The study conducted within the Barcelonnette basin defines rainfall patterns associated to fast-gravitational movement (e.g. debris flows) and slow-gravitational movement (e.g. mudslides): 1. At the monthly scale, a clear distinction can be observed for debris flows (strong
storm associated to a 7-day dry period) and mudslide (at least 40 mm of antecedent precipitation during the last 7 days and at least 200 mm during the last 180 days)
2. At the daily scale, the correlation between events and the rainfall of the triggering date is quite good, no threshold can be established due to the inaccurate knowledge on rainfall variability (e.g. the upper part of the hillslopes)
3. At the hourly scale, debris flows triggering is associated to relatively short and intense summer thunderstorms while mudslides are associated to longer rainfall sequences exhibiting low average and peak intensities.
This work is being continued in order to define potential indicators for the Waidhofen/Ybbs study area.
Indicators of changes: hydro-climatic
RESULTS
PROGRESS
Indicators of changes: hydro-climatic
Indicators of changes: Geovisualization
For the study area Waidhofen/Ybbs it is was not possible to implement a demonstration platform due to data usage restrictions. This was agreed on in the project meeting in Vienna, June 2011 with the consortium. However UNIVIE was regularly involved in the development of the demonstration platform Barcelon@.
Barcelon@ is a WebGIS application based on open source standards. It supplies
a system to decrease the disparity between scientific results and stakeholders’ practical needs (simple interface, easy-to-use buttons in a generally user-friendly approach). Secondly the wide collection of assorted information (e.g. landslide controlling factors, susceptibility maps, information on elements at risk and their vulnerability, administrative data) and the data comparison can offer a feasible support in the decision-making process.
The application is based on different levels of access (citizens or end-users) and completely supports Geo Web Services (useful for external connections).
RESULTS
RESULTS
Demonstration Platform and Geovisualization
Data Storage. A huge amount of data is achieved. Resample, compression and clipping (based on municipality level) of every layers arranged a homogeneous geo-database all over the entire Basin
RESULTS
Cartoserver frame. Data collected are transferred inside a mapfile (mrisk.map). The aim is to fix classification, transparency, location and visual rules of all the dataset. Thus the spatial geoinformation is adapted in framework standards
Demonstration Platform and Geovisualization
RESULTS
Cartoclient frame. Every client request is standardized. The plugins to export information (PDF, CSV files, layout template) and final browser visualization are managed here. The web services guarantee connection with Cartoserver for every front-end request by users
Demonstration Platform and Geovisualization
PROGRESS
The database gathers information from different institutes and agencies, but it is standardized for scale, classification and resolution. Thus user can have a homogenous “state of art” of all the available information necessary in risk management, considering the original accuracy of dataset involved. Different tasks are available inside Barcelon@ considering geo-spatial tools and all information covering the study area. The criteria to create different clusters supply the proof of scientific results performed and the state-of-art about knowledge on natural events
The wide collection of assorted information and the data comparison offers a
great support in the decision-making process. CartoWeb is the open source type of platform selected for the task. It is a comprehensive and ready-to-use WebGIS and supplies advanced development. The architecture of the service is totally customized through visualization needs and the task of the research
The service is currently under test and will be online by end 2013