CMCC Use Case Presentation Edoardo Bucchignani, Paola Mercogliano - PDF document

13/03/2013 CMCC Use Case Presentation Edoardo Bucchignani, Paola Mercogliano and ISC team Euromediterranean Center on Climate Change Impact on Soil and Coasts Division Use Case defined in JRA5 Evaluation of the effects of climate changes on landslides phenomena triggered by precipitation events (clay strain). The landslide mass moves with considerable internal deformations; at the base of the landslide a flow surface exists. How will the ongoing climate change influence the landslide activity in the Mediterranean area (and in special way in Italy) ? 2 1

13/03/2013 Use Case Description Summary : this test case describes the procedure used in order to make usable the output data of the regional climate model COSMO ‐ CLM to the SEEP/W (GEO ‐ SLOPE International Ltd.) model. The last one is used for stability analysis in order to evaluate landslide risk scenarios. Location: Bucciano, S. Martino V. C., Cervinara (Italy). Data needs: atmospheric variables on mesoscale (obtained using the COSMO ‐ CLM regional model initialized by the global model CMCC ‐ MED): precipitation, temperature at 2m (daily, min and max) wind at 10 m , relative humidity. Resolution: 8 km , every 6 hours. 3 Use Case Description • CMCC climate researchers prepare datasets in NETCDF format containing the solutions (saved every 6 hours) related to a well defined domain (i.e. Italy) for past period (1971-2000) and XXI century (A1B scenario). • These data are stored on a Data Center. • Geotechnical engineers search on these datasets the information related to the subdomain and to the period of their interest. If this subdomain contains more than one grid points, data are averaged in order to produce a single value for each variable (at each time step). • Geotechnical Engineers download data via FTP or WEB. • Geotechnical Engineers run their Finite Elements model (SEEP/W). 4 2

13/03/2013 Details on the Num. Hydro. Climate Chain Output COSMO-CLM NETCDF Files: each file contains the solution saved every 6 hrs, with a resolution of 8 km. (T2m, Totprec, U10m, rel hum…) Postprocessing (UNIX scripts based on CDO) Ascii files containing the transient history of variables (e.g. T2M and Tot. Prec.) in the site of interest SEEP/W: Finite Element Code for the evaluation of the effects on the soil of precipitations 5 The COSMO ‐ CLM regional climate model • The COSMO ‐ CLM is the Climate Mode of the COSMO model system: • It is a non hydrostatic regional climate model atmospheric prediction system, developed by the CLM ‐ Community. •It is designed for simulations on time scales up to centuries and spatial resolutions down to 1 km. • It is the only limited area numerical model system in Europe which has a range of applicability encompassing: 1. operational numerical weather prediction (COSMO) 2. regional climate modelling of past, present and future (CLM), 3. the dispersion of trace gases and aerosol (ART) and 4. idealized studies (ITC) • It is applicable for downscaling in all regions of the world and of most of the Global Climate simulations available • It is fully documented. 3

13/03/2013 Climate change in Mediterranean area Giorgi and Lionello (2008) • decreasing yearly cumulated rainfall values: 2001 ‐ 2020 2081 ‐ 2100 summer ‐ 7% ‐ 28% fall ‐ 3% ‐ 15% spring ‐ 2% ‐ 14% winter ‐ 2% ‐ 8% • increasing surface warming: 2001 ‐ 2020 2081 ‐ 2100 summer +0.7°C +4.6°C winter +0.7°C +3.1°C • intensified hydrological cycle due to greater atmospheric water holding capacity • increasing variability: heavier rainfall events separated by longer dry periods 7 Landslides with a shorter antecedent period, or those exhibit a strong response to relatively short term rainfall events, could be more vulnerable to the predicted increase variability of winter rainfall and so could experience a shorter return period for slope movements. …However, slope instability caused by longer periods of antecedent rainfall may occur less often in response to drier summers with increased evapotranspiration” Dixon&Brook(2007 )) 8 4

13/03/2013 Previous works MamTor(England) Boisivre (SE France) the Roughs (SE England) case study effective precipitation ‐ > effective precipitation ‐ > effective precipitation ‐ > 180 days wet season rainfalls 90 ‐ 120 days Dixon&Brook(2006 ) Buma&Dehn (1998;1999; Collison et al. (2000 ) 2000) GCM (HADCM3)+ GCM (HADCM2GG) GCM (ECHAM4 ‐ OPYC3; nested models until emission scenario HADCM2GG; 50km resolution climatological IPCCIS92a HADCM2GS) model emission scenario IPCCIS92a Direct GCM interpolation; weather generator multiple regression; downscaling analog ‐ downscaling approach empirical rainfall thresholds; 1D hydrological model rainfall thresholds bucket models; +GIS slope stability slope model (hydrological analysis balance+ infinite slope 9 impact model approach) CMCC first attempt: quantile based bias correction Basento valley (Basilicata Region, South Italy) • case study • effective precipitation ‐ > wet season rainfalls • Comegna et al. (Landslides, 2012, DOI: 10.1007/s10346 ‐ 012 ‐ 0339 ‐ 3) climatological COSMO ‐ CLM 4.8 forced by CMCC ‐ MED; spatial resolution 8 km; IPCC scenario A1B; time period 1965 ‐ 2100 model «quantile based» bias correction approach (Wood et al., 2004) downscaling approach hydrological : Vadose solving numerically the simoultaneous equations regulating thermal and hydraulic soil behaviour impact model mechanical : relationship between pore water pressure and landslide movements 1 0 5

13/03/2013 Main results: expected (water pressure) landslide velocities 1 1 CMCC second attempt: intermediate degree of grain • Cervinara (Campania Region, south Italy) case study • effective precipitation (coupled effect of a trigger precipitation over a short interval [24ore] + antecedent precipitations over a long time span [90d] • Damiano & Mercogliano [2011] (II World Landslides Forum ) COSMO ‐ CLM 4.8 forced by CMCC ‐ MED; spatial resolution 8 km; IPCC scenario climatological A1B; considered time 1965 ‐ 2100 model • increase in storm intensities and number of dry days; no change in yearly rainfall assumptions • increase in storm intensities during winter • heavy rainfall event added at the end of wet season based on expected changes in annual rainfall regime IMOD 3D ‐ GIS model regional scale home ‐ made code hydrological : solving numerically Richards’ equation for estimating water pressure in isothermal conditions impact model 1 mechanical : infinite slope approach 2 6

13/03/2013 Main results: comparison between factors of stability (FS) �� ����� � ��� � ���� ����� �������� � � = ����� ������ = ���������� Assumption 1: increase in storm intensities and number of dry days; no change in yearly rainfall Assumption 2: increase in storm intensities during winter Assumption 3: heavy rainfall event added at the end of wet season 1 3 Under development Step 1 : Simulation of the climate of the period 1971 ‐ 2000 with the regional model COSMO ‐ CLM forced by ERA40 as perfect boundary conditions. Data are calibrated using the MOS downscaling technique. Step 2 : Simulation of the climate of the period 1971 ‐ 2000 with the regional model COSMO ‐ CLM forced by a global model. Calibration as in step 1. Step 3 : Generation of regional climate scenarios for the XXI century . Calibration as in step 1  Input for the impact numerical modelization. 1 4 7

13/03/2013 Under development MOS analog method: ‐ the closest historical predictor B (the analog) considering the Euclidean distance between the two raw predictor fields is found, ‐ then, the local precipitation observed, b, correspondent to the analog day B, is used as the downscaled precipitation of the day A (i.e. a=b). Turco et al. (2011) MOS downscaling for hydrological application 1 5 Under development CASE ‐ HISTORIES ‐ Orvieto (Umbria,Central Italy) ‐ Fosso San Martino (Abruzzo,Central Italy) ‐ Cervinara (Campania, South Italy) they are selected by virtue of wide availability of data regarding previous landslide activity, soil hydraulic and mechanical behaviour, meteorological observations 1 6 8

13/03/2013 Weather Analyst Gis tool for the study of the impact on soil of climate changes All the impact studies request data easily manipulated by GIS infrastructures and high space/time resolutions Weather Analyst is an extension of Arcgis Desktop able to manage atmospheric data provided by different sources (climate models, satellite data, radar, observations) The tools allows to perform: • Analysis of historical series • Analysis of climate scenarios • Statistical and stochastical downscaling 1 7 Interface Scheme Input Weather Analyst Griglie Cosmo LM Dati Satellitari NetCDF Catalogo Export Output IMOD ArcMap / Esri tools Esri Xls .M Files .Shp / .Kml GeoDB Web GIS Rainfall Data 9

13/03/2013 Downscaling techniques implemented Not in literature Not in literature RainFARM RainFARM MRI MRI Kriging Kriging RBF RBF Spline Spline IDW IDW Probabilistic Probabilistic Natural Natural Neighborhood Neighborhood Neighbor Neighbor Trend Trend Polynomial Polynomial Primo vicino Primo vicino Con buffer Con buffer Primo vicino Primo vicino Interpolating Interpolating Thanks 10