H41G-0842
A Validation of Remotely Sensed Fires Using Ground Reports
A satellite based analysis of fire detections and smoke emissions for North America is produced daily by NOAA/NESDIS. The analysis incorporates data from the MODIS (Terra and Aqua) and AVHRR (NOAA-15/16/17) polar orbiting instruments and GOES East and West geostationary spacecraft with nominal resolutions of 1km and 4 km for the polar and geostationary platforms respectively. Automated fire detection algorithms are utilized for each of the sensors. Analysts perform a quality control procedure on the automated detects by deleting points that are deemed to be false detects and adding points that the algorithms did not detect. A limited validation of the final quality controlled product was performed using high resolution (30 m) ASTER data in the summer of 2006. Some limitations in using ASTER data are that each scene is only approximately 3600 square km, the data acquisition time is relatively constant at around 1030 local solar time and ASTER is another remotely sensed data source. This study expands on the ASTER validation by using ground reports of prescribed burns in Montana and Idaho for 2003 and 2004. It provides a non-remote sensing data source for comparison. While the ground data do not have the limitations noted above for ASTER there are still limitations. For example, even though the data set covers a much larger area (nearly 600,000 square km) than even several ASTER scenes, it still represents a single region of North America. And while the ground data are not restricted to a narrow time window, only a date is provided with each report, limiting the ability to make detailed conclusions about the detection capabilities for specific instruments, especially for the less temporally frequent polar orbiting MODIS and AVHRR sensors. Comparison of the ground data reports to the quality controlled fire analysis revealed a low rate of overall detection of 23.00% over the entire study period. Examination of the daily detection rates revealed a wide variation, with some days resulting in as little as 5 detects out of 107 reported fires while other days had as many as 84 detections out of 160 reports. Inspection of the satellite imagery from the days with very low detection rates revealed that extensive cloud cover prohibited satellite fire detection. On days when cloud cover was at a minimum, detection rates were substantially higher. An estimate of the fire size was also provided with the ground data set. Statistics will be presented for days with minimal cloud cover which will indicate the probability of detection for fires of various sizes. http://www.ssd.noaa.gov/PS/FIRE/hms.html
H41G-0843
The dynamics of active landslide development and evolution: a combined structural geology, geomorphology and InSAR approach.
In recent years structural geology has been used as a tool to investigate the development and evolution of potential rockslides. Recent studies have been mostly concentrated on identifying particular geometrical constellations suitable for sliding to occur and the observed kinematics. Limited emphasis on the direct relationships between the development of structures, evidence for movement, and its effect on the geomorphological architecture have been described. A reconciliation between field observations and various standard measuring techniques has often proven ambiguous or problematic. However, recent technological advances in interferometric synthetic aperture radar (InSAR) satellite technology provide a new measurement method to determine potential rockslide movement and therefore provide a direct link between qualitative movement data and field observations of structures, kinematics and geomorphological change in slope. We present structural and geomorphological observations from the Gamanjunni slide in Troms, Norway, combined with detailed InSAR deformation measurement data based on analysis of ERS-1/2 SAR data in the 1992--1999 timeframe, to determine the detailed activity, evolution and detailed magnitude of displacement. The slide is located on a west-facing mountainside at a height of 1200m and is made up of two angled back-scarps with a 20-30° basal sliding plane which outcrops at the front of the moving block. On a locality scale we can demonstrate that InSAR documents active landslide movement which is in agreement with several different observation types in the field. The spatial extent of the InSAR movement area also fits extremely well with the area delimited by active structures. In detail at a sub-locality scale, InSAR is also able to document structural processes in the moving block. Oblique kinematics means that the two angled fault scarps have different movement rates. This is reflected in the observed, differential activity of the two fault scarps which is directly reflected in a different movement magnitude in the InSAR data. We also document a segmental, partial reactivation of one of the scarps which is reflected in the InSAR data as an along strike variation in InSAR movement. Field evidence suggests that some fault segments have been active at different times and that previously active fault segments, which are now dead, have been superceded by newer, more active faults which are now accommodating present movement. This is also reflected in the details of the InSAR data. Therefore we conclude that InSAR is an essential tool in providing quantitative data of the activity in landslides and an important confirmation of the field observations for recent movement activity. However, we go further in suggesting that the InSAR greatly extends our knowledge on the processes and evolution of fault scarp development and therefore active slide evolution.
H41G-0844
Landslide Forecasting Using Microwave Remote Sensing
Landslides, a natural disaster, are becoming more common in mountainous region of many countries. A dynamic physically-based slope stability model that requires soil moisture or wetness index can be driven by remote sensing products from multiple Earth observing platforms. While satellite remote sensing (e.g., AMSR-E and TRMM satellite data) can measure surface soil moisture, land surface model can estimate the soil moisture profile. This research compares AMSR-E surface soil moisture with the variable infiltration capacity (VIC-3L) model's soil moisture at Cleveland Corral landslide area in California, USA from 2003 to 2006. Snow cover influences on AMSR-E surface soil moisture estimate are also examined. The results show a strong relationship among AMSR-E's surface soil moisture, modeled soil moisture and in situ pore water pressure measurements. Preliminary results show that the AMSR-E satellite data, coupled with VIC model estimates, are viable for rainfall induced slope stability analysis at regional or global scales. Keywords: Landslide, AMSR-E, TRMM, Remote Sensing, VIC, Soil Moisture
H41G-0845
A New Approach to Liquefaction Potential Mapping Using Remote Sensing and Machine Learning
In order to help communities better plan and mitigate the effects of seismic hazards, it is important to use innovations in science and technology to improve our techniques for mapping the spatial extents of seismic hazards. Earthquake induced ground shaking in areas with saturated sandy soils pose a major threat to communities as a result of the soil liquefaction. Liquefaction is the process of changing a saturated cohesionless soil from a solid to liquid state due to increased pore pressure. Many major earthquakes, especially those in coastal regions, result in liquefaction related ground failures that can lead to infrastructure damage or slope stability issues. Currently liquefaction potential is assessed on two scales: regionally based on surficial geologic unit or locally based on geotechnical sample data. Regional liquefaction potential maps fail to capture the variability of liquefaction potential on the local scale. On the other hand, collection of geotechnical data on the local scale is costly and only done for specific engineering projects and therefore not generally available for regional mapping. Today, the advent of advanced remote sensing products from air and space borne sensors allow us to explore the land surface parameters (geology, moisture content, temperature) at different spatial scales (remote sensor footprint). In this study, we explore the use of satellite based remote sensing data (Landsat 7 ETM+), together with digital elevation model, ground water table, land cover classification, geology, water index and normalized difference vegetation index (NDVI) to characterize the liquefaction potential of northern Monterey and southern Santa Cruz counties in California. A supervised classification of the data into seven classes based on the liquefaction potential map developed by Dupre and Tinsley 1980 was done using Support Vector Machine (SVM). SVM is a machine learning/artificial intelligence algorithm that has the ability to simulate the learning capabilities of a human brain and make appropriate predictions that involve intuitive judgments and a high degree of nonlinearity. The accuracy of the developed liquefaction potential map was tested using independent testing data that was not used for the model development. The results show that the developed liquefaction potential map has an overall classification accuracy of 84%, indicating that the combination of remote sensing data and other relevant spatial data together with machine learning can be a promising approach for liquefaction potential mapping.
H41G-0846
Hurricane related flooding monitoring: a method to delineate potentially affected areas by using a GIS model in the Caribbean area
This research integrates the concept that the subject of natural hazards and the use of existing remote sensing systems in the different phases of a disaster management for a specific hurricane hazard, is based on the applicability of GIS model for increasing preparedness and providing early warning. The modelling of an hurricane event in potentially affected areas by GIS has recently become a major topic of research. In this context the disastrous effects of hurricanes on coastal communities and surroundings areas are well known, but there is a need to better understand the causes and the hazards contributions of the different events related to an hurricane, like storm surge, flooding and high winds. This blend formed the basis of a semi- quantitative and promising approach in order to model the spatial distribution of the final hazard along the affected areas. The applied model determines a sudden onset zoning from a set of available parameters starting from topography based on Shuttle Radar Topography Mission (SRTM) data. From the Digital Elevation Model as a first step the river network is derived and then classified based on the Strahler order account as proportional to flooding area. Then we use a hydrologic model that uses the wetness index (a parameter of specific catchment area defined as upslope area per unit contour length) to better quantify the drainage area that contributes to the flooded events. Complementary data for the final model includes remote sensed density rain dataset for the hurricane events taking into account and existing hurricane tracks inventories together with hurricane structure model (different buffers related to wind speed hurricane parameters in a GIS environment). To assess the overall susceptibility, the hazard results were overlaid with population dataset and landcover. The approach, which made use of a number of available global data sets, was then validated on a regional basis using past experience on hurricane frequency study over an area that covers both developed and developing countries in the Caribbean region. As a final result we can state that remote sensing data analysed together with meteorological and environmental data in an integrated GIS system give a spatially resolved picture of the surface conditions and, in our context, informations on the occurrence, extent and severity of hurricane hazard. The applied GIS model has then given rise to a long-lead system that can be set-up to allow such a early warning to go ahead.
H41G-0847
Monitoring Of Soil Erosion Using Digital Camera Under Simulated Rainfall
The photogrammetry is often used to measure the amount of erosion from Digital Elevation Model (DEM). A set of DEM data for different periods is particularly important for soil erosion measurement. It helps in evaluating surface roughness changes. The photogrammetry application to soil erosion has been studied for the purpose of river channel shifting, landslide movement and gully erosion. However, there is a lack of information in monitoring small scale of soil erosion like sheet erosion, which is a slight change in soil surface compared to gully. Due to the limitation in the accuracy of photogrammetry, the objective of this study was to assess the precision of photogrammetry. A new system equipped with an automated photogrammetry was developed in our laboratory. The system consists of two digital cameras, a rain simulator with a twelve meters high tower, and a set computer image analysis program of three-dimensional algorithm. The precision was first assessed by comparing the measured data against photogrammetry under no rainfall condition. The assessment was based on the statistical parameters, the absolute error (AE) and relative error (RE), which are 0.0946 kg/m2 and 8.59%, respectively. In the second step, the precision was assessed for under simulated rainfall on soil surface packed to two different dry bulk densities. At 1.20 g/cm3 density, the AE and RE were 6.62 kg/m2 and 1490%, respectively and corresponding values at 1.30 g/cm3 density were 0.675 kg/m2 and 86.0%, respectively. The soil saturated with water and whose bulk density was 1.38 g/cm3, was subjected to rainfall of intensities of 40, 80, 120 mm/h continuously for 1 hour. They had AE (RE) of 0.0313 (36.1%), 0.0183 (26.4%), 0.0268 kg/m2 (29.8%), respectively, at each intensity. It was concluded that is possible to measure soil sheet erosion of small scale with high precision and under the rainfall.