To download PDF of paper, click on ‘Links’ underneath citation, then click the URL listed.
2016 |
Lee, Jun-Hak; Biging, Gregory; Fisher, Joshua An Individual Tree-Based Automated Registration of Aerial Images to Lidar Data in a Forested Area Journal Article Photogrammetric Engineering & Remote Sensing, 82 (9), pp. 699-701, 2016. Abstract | Links | BibTeX | Tags: aerial images, LiDAR, tree tops @article{Jun-HakLee2016, title = {An Individual Tree-Based Automated Registration of Aerial Images to Lidar Data in a Forested Area}, author = {Jun-Hak Lee and Gregory Biging and Joshua Fisher}, doi = {https://doi.org/10.1016/S0099-1112(16)30121-5}, year = {2016}, date = {2016-08-01}, journal = {Photogrammetric Engineering & Remote Sensing}, volume = {82}, number = {9}, pages = {699-701}, abstract = {In this paper, we demonstrate an approach to align aerial images to airborne lidar data by using common object features (tree tops) from both data sets under the condition that conventional correlation-based approaches are challenging due to the fact that the spatial pattern of pixel gray-scale values in aerial images hardly exist in lidar data. We extracted tree tops by using an image processing technique called extended-maxima transformation from both aerial images and lidar data. Our approach was tested at the Angelo Coast Range Reserve on the South Fork Eel River forests in Mendocino County, California. Although the aerial images were acquired simultaneously with the lidar data, the images had only approximate exposure point locations and average flight elevation information, which mimicked the condition of limited information availability about the aerial images. Our results showed that this approach enabled us to align aerial images to airborne lidar data at the single-tree level with reasonable accuracy. With a local transformation model (piecewise linear model), the rmse and the median absolute deviation (mad) of the registration were 9.2 pixels (2.3 meters) and 6.8 pixels (1.41 meters), respectively. We expect our approach to be applicable to fine scale change detection for forest ecosystems and may serve to extract detailed forest biophysical parameters. }, keywords = {aerial images, LiDAR, tree tops}, pubstate = {published}, tppubtype = {article} } In this paper, we demonstrate an approach to align aerial images to airborne lidar data by using common object features (tree tops) from both data sets under the condition that conventional correlation-based approaches are challenging due to the fact that the spatial pattern of pixel gray-scale values in aerial images hardly exist in lidar data. We extracted tree tops by using an image processing technique called extended-maxima transformation from both aerial images and lidar data. Our approach was tested at the Angelo Coast Range Reserve on the South Fork Eel River forests in Mendocino County, California. Although the aerial images were acquired simultaneously with the lidar data, the images had only approximate exposure point locations and average flight elevation information, which mimicked the condition of limited information availability about the aerial images. Our results showed that this approach enabled us to align aerial images to airborne lidar data at the single-tree level with reasonable accuracy. With a local transformation model (piecewise linear model), the rmse and the median absolute deviation (mad) of the registration were 9.2 pixels (2.3 meters) and 6.8 pixels (1.41 meters), respectively. We expect our approach to be applicable to fine scale change detection for forest ecosystems and may serve to extract detailed forest biophysical parameters. |
2014 |
Bode, Collin; Limm, Michael; Power, Mary; Finlay, Jacques Subcanopy Solar Radiation model: Predicting solar radiation across a heavily vegetated landscape using LiDAR and GIS solar radiation models Journal Article Remote Sensing of Environment, 154 (SI), pp. 387-397, 2014, ISSN: 0034-4257. Abstract | Links | BibTeX | Tags: insolation, LiDAR, solar model, subcanopy, vegetation shading @article{Bode2014, title = {Subcanopy Solar Radiation model: Predicting solar radiation across a heavily vegetated landscape using LiDAR and GIS solar radiation models}, author = {Collin Bode and Michael Limm and Mary Power and Jacques Finlay}, url = {https://angelo.berkeley.edu/wp-content/uploads/2015/01/Bode_SSR_Methods_RSE2013.pdf}, doi = {DOI:10.1016/j.rse.2014.01.028}, issn = {0034-4257}, year = {2014}, date = {2014-11-14}, journal = {Remote Sensing of Environment}, volume = {154}, number = {SI}, pages = {387-397}, abstract = {Solar radiation flux, irradiance, is a fundamental driver of almost all hydrological and biological processes. Ecological models of these processes often require data at the watershed scale. GIS-based solar models that predict insolation at the watershed scale take topographic shading into account, but do not account for vegetative shading. Most methods that quantify subcanopy insolation do so only at a single point. Further, subcanopy model calibration requires significant field effort and knowledge of characteristics (species composition, leaf area index & mean leaf angle for each species), and upscaling to watersheds is a significant source of uncertainty. We propose an approach to modeling insolation that uses airborne LiDAR data to estimate canopy openness as a Light Penetration Index (LPI). We couple LPI with the GRASS GIS r.sun solar model to produce the Subcanopy Solar Radiation model (SSR). SSR accounts for both topographic shading and vegetative shading at a landscape scale. After calibrating the r.sun model to a weather station at our study site, we compare SSR model predictions to black thermopile pyranometer field measurements and to hemispherical photographs using Gap Light Analyzer software, a standard method for point estimation of subcanopy radiation. Both SSR and hemispherical models exhibit a similar linear relationship with pyranometer data, and the models predict similar total solar radiation flux across the range of canopy openness. This approach allows prediction of light regimes at watershed scales with resolution that was previously possible only for local point measurements. }, keywords = {insolation, LiDAR, solar model, subcanopy, vegetation shading}, pubstate = {published}, tppubtype = {article} } Solar radiation flux, irradiance, is a fundamental driver of almost all hydrological and biological processes. Ecological models of these processes often require data at the watershed scale. GIS-based solar models that predict insolation at the watershed scale take topographic shading into account, but do not account for vegetative shading. Most methods that quantify subcanopy insolation do so only at a single point. Further, subcanopy model calibration requires significant field effort and knowledge of characteristics (species composition, leaf area index & mean leaf angle for each species), and upscaling to watersheds is a significant source of uncertainty. We propose an approach to modeling insolation that uses airborne LiDAR data to estimate canopy openness as a Light Penetration Index (LPI). We couple LPI with the GRASS GIS r.sun solar model to produce the Subcanopy Solar Radiation model (SSR). SSR accounts for both topographic shading and vegetative shading at a landscape scale. After calibrating the r.sun model to a weather station at our study site, we compare SSR model predictions to black thermopile pyranometer field measurements and to hemispherical photographs using Gap Light Analyzer software, a standard method for point estimation of subcanopy radiation. Both SSR and hemispherical models exhibit a similar linear relationship with pyranometer data, and the models predict similar total solar radiation flux across the range of canopy openness. This approach allows prediction of light regimes at watershed scales with resolution that was previously possible only for local point measurements. |
2013 |
Handwerger, Alexander L; Roering, Joshua J; Schmidt, David A Controls on the seasonal deformation of slow-moving landslides Journal Article Earth and Planetary Science Letters, 377-378 , pp. 239–247, 2013. Abstract | Links | BibTeX | Tags: hydrology, InSAR, landslides, LiDAR, pore-water pressure diffusion, precipitation @article{Handwergera2013, title = {Controls on the seasonal deformation of slow-moving landslides}, author = {Alexander L. Handwerger and Joshua J. Roering and David A. Schmidt}, url = {https://angelo.berkeley.edu/wp-content/uploads/Handwerger_2013_EarthPlanSciLetters.pdf}, doi = {10.1016/j.epsl.2013.06.047}, year = {2013}, date = {2013-09-01}, journal = {Earth and Planetary Science Letters}, volume = {377-378}, pages = {239–247}, abstract = {Precipitation drives seasonal velocity changes in slow-moving landslides by increasing pore-water pressure and reducing the effective normal stress along basal shear zones. This pressure change is often modeled as a pore-water pressure wave that diffuses through the landslide body, such that the minimum time required for landslides to respond to rainfall should vary as the square of landslide depth (which often approximates the saturated thickness) and inversely with hydraulic diffusivity. Here, we assess this model with new observations from the landslide-prone Eel River catchment, Northern California. Using satellite radar interferometry (InSAR) time series, precipitation data, and high-resolution topographic data from airborne lidar, we quantify the seasonal dynamics of 10 slow-moving landslides, which share the same lithologic, tectonic, and Mediterranean climate conditions. These slope failures have areas ranging from 0.16 to 3.1 km2, depths that vary from 8 to 40 m, and average downslope velocities of 0.2 to 1.2 m/yr. Each slide exhibits well-defined seasonal velocity changes with a periodicity of ∼1 yr and responds (i.e., accelerates) within 40 days following the onset of rainfall. Despite a five-fold variation in landslide depth, we do not detect systematic differences in response time within the resolution of our observations. Our results could imply that: 1) slides in our study area are sensitive to subtle hydrologic perturbations, 2) the ‘effective’ diffusivity governing slide behavior is much higher than field-derived values because pore pressure transmission and slide dynamics are facilitated by preferential flow paths, particularly cracks related to deformation and seasonal shrink-swell cycles, or 3) a simple one-dimensional linear diffusion model may fail to capture the three-dimensional time-dependent hydrologic changes inherent in an evolving mechanical–hydrologic system, such as a slow-moving landslide.}, keywords = {hydrology, InSAR, landslides, LiDAR, pore-water pressure diffusion, precipitation}, pubstate = {published}, tppubtype = {article} } Precipitation drives seasonal velocity changes in slow-moving landslides by increasing pore-water pressure and reducing the effective normal stress along basal shear zones. This pressure change is often modeled as a pore-water pressure wave that diffuses through the landslide body, such that the minimum time required for landslides to respond to rainfall should vary as the square of landslide depth (which often approximates the saturated thickness) and inversely with hydraulic diffusivity. Here, we assess this model with new observations from the landslide-prone Eel River catchment, Northern California. Using satellite radar interferometry (InSAR) time series, precipitation data, and high-resolution topographic data from airborne lidar, we quantify the seasonal dynamics of 10 slow-moving landslides, which share the same lithologic, tectonic, and Mediterranean climate conditions. These slope failures have areas ranging from 0.16 to 3.1 km2, depths that vary from 8 to 40 m, and average downslope velocities of 0.2 to 1.2 m/yr. Each slide exhibits well-defined seasonal velocity changes with a periodicity of ∼1 yr and responds (i.e., accelerates) within 40 days following the onset of rainfall. Despite a five-fold variation in landslide depth, we do not detect systematic differences in response time within the resolution of our observations. Our results could imply that: 1) slides in our study area are sensitive to subtle hydrologic perturbations, 2) the ‘effective’ diffusivity governing slide behavior is much higher than field-derived values because pore pressure transmission and slide dynamics are facilitated by preferential flow paths, particularly cracks related to deformation and seasonal shrink-swell cycles, or 3) a simple one-dimensional linear diffusion model may fail to capture the three-dimensional time-dependent hydrologic changes inherent in an evolving mechanical–hydrologic system, such as a slow-moving landslide. |
2011 |
Mackey, Benjamin H; Roering, Joshua J Geological Society of America Bulletin, 123 (7-8), pp. 1560-1576, 2011. Abstract | Links | BibTeX | Tags: aerial photographs, earthflows, LiDAR @article{Mackey2011, title = {Sediment yield, spatial characteristics, and the long-term evolution of active earthflows determined from airborne LiDAR and historical aerial photographs, Eel River, California}, author = {Benjamin H. Mackey and Joshua J. Roering}, url = {https://angelo.berkeley.edu/wp-content/uploads/Mackey_2011_GeolSociAmerBul.pdf}, doi = {10.1130/B30306.1}, year = {2011}, date = {2011-01-01}, journal = {Geological Society of America Bulletin}, volume = {123}, number = {7-8}, pages = {1560-1576}, abstract = {In mountainous landscapes with weak, fine-grained rocks, earthflows can dominate erosion and landscape evolution by supplying sediment to channels and controlling hillslope morphology. To estimate the contribution of earthflows to regional sediment budgets and identify patterns of landslide activity, earthflow movement needs to be quantified over significant spatial and temporal scales. Presently, there is a paucity of data that can be used to predict earthflow behavior beyond the seasonal scale or over spatially extensive study areas. Across 226 km2 of rapidly eroding Franciscan Complex rocks of the Eel River catchment, northern California, we used a combination of LiDAR (light detection and ranging) and orthorectified historical aerial photographs to objectively map earthflow movement between 1944 and 2006. By tracking the displacement of trees growing on earthflow surfaces, we find that 7.3% of the study area experienced movement over this 62 yr interval, preferentially in sheared argillaceous lithology. This movement is distributed across 122 earthflow features that have intricate, elongate planform shapes, a preferred south-southwesterly aspect, and a mean longitudinal slope of 31%. The distribution of mapped earthflow areas is well-approximated by a lognormal distribution with a median size of 36,500 m2. Approximately 6% of the study area is composed of earthflows that connect to major channels; these flows generated an average sediment yield of 19,000 t km−2 yr−1 (rock erosion rate of ∼7.6 mm/yr) over the 62 yr study period, equating to a regional yield of 1100 t km−2 yr−1 (∼0.45 mm/yr) if distributed across the study area. As such, a small fraction of the landscape can account for half of the regional denudation rate estimated from suspended sediment records (2200 t km−2 yr−1 or ∼0.9 mm/yr). We propose a conceptual model for long-term earthflow evolution wherein earthflows experience intermittent activity and long periods of dormancy when limited by the availability of readily mobilized sediment on upper slopes. Ultimately, high-order river channels and ephemeral gully networks may serve to destabilize hillslopes, controlling the evolution of earthflow-prone terrain. }, keywords = {aerial photographs, earthflows, LiDAR}, pubstate = {published}, tppubtype = {article} } In mountainous landscapes with weak, fine-grained rocks, earthflows can dominate erosion and landscape evolution by supplying sediment to channels and controlling hillslope morphology. To estimate the contribution of earthflows to regional sediment budgets and identify patterns of landslide activity, earthflow movement needs to be quantified over significant spatial and temporal scales. Presently, there is a paucity of data that can be used to predict earthflow behavior beyond the seasonal scale or over spatially extensive study areas. Across 226 km2 of rapidly eroding Franciscan Complex rocks of the Eel River catchment, northern California, we used a combination of LiDAR (light detection and ranging) and orthorectified historical aerial photographs to objectively map earthflow movement between 1944 and 2006. By tracking the displacement of trees growing on earthflow surfaces, we find that 7.3% of the study area experienced movement over this 62 yr interval, preferentially in sheared argillaceous lithology. This movement is distributed across 122 earthflow features that have intricate, elongate planform shapes, a preferred south-southwesterly aspect, and a mean longitudinal slope of 31%. The distribution of mapped earthflow areas is well-approximated by a lognormal distribution with a median size of 36,500 m2. Approximately 6% of the study area is composed of earthflows that connect to major channels; these flows generated an average sediment yield of 19,000 t km−2 yr−1 (rock erosion rate of ∼7.6 mm/yr) over the 62 yr study period, equating to a regional yield of 1100 t km−2 yr−1 (∼0.45 mm/yr) if distributed across the study area. As such, a small fraction of the landscape can account for half of the regional denudation rate estimated from suspended sediment records (2200 t km−2 yr−1 or ∼0.9 mm/yr). We propose a conceptual model for long-term earthflow evolution wherein earthflows experience intermittent activity and long periods of dormancy when limited by the availability of readily mobilized sediment on upper slopes. Ultimately, high-order river channels and ephemeral gully networks may serve to destabilize hillslopes, controlling the evolution of earthflow-prone terrain. |
2010 |
Passalacqua, Paola; Trung, Tien Do; Foufoula-Georgiou, Efi; Sapiro, Guillermo; Dietrich, William E A geometric framework for channel network extraction from lidar: Nonlinear diffusion and geodesic paths Journal Article Journal of Geophysical Research-Earth Surface, 115 (F1), 2010. Abstract | Links | BibTeX | Tags: Channel network, geodesic paths, LiDAR, nonliner diffusion @article{Passalacqua2010, title = {A geometric framework for channel network extraction from lidar: Nonlinear diffusion and geodesic paths}, author = {Paola Passalacqua and Tien Do Trung and Efi Foufoula-Georgiou and Guillermo Sapiro and William E. Dietrich}, url = {https://angelo.berkeley.edu/wp-content/uploads/Passalacqua_2010_JoGeophysRes.pdf}, doi = {10.1029/2009JF001254}, year = {2010}, date = {2010-01-07}, journal = {Journal of Geophysical Research-Earth Surface}, volume = {115}, number = {F1}, abstract = {A geometric framework for the automatic extraction of channels and channel networks from high-resolution digital elevation data is introduced in this paper. The proposed approach incorporates nonlinear diffusion for the preprocessing of the data, both to remove noise and to enhance features that are critical to the network extraction. Following this preprocessing, channels are defined as curves of minimal effort, or geodesics, where the effort is measured on the basis of fundamental geomorphological characteristics such as flow accumulation area and isoheight contours curvature. The merits of the proposed methodology, and especially the computational efficiency and accurate localization of the extracted channels, are demonstrated using light detection and ranging (lidar) data of the Skunk Creek, a tributary of the South Fork Eel River basin in northern California.}, keywords = {Channel network, geodesic paths, LiDAR, nonliner diffusion}, pubstate = {published}, tppubtype = {article} } A geometric framework for the automatic extraction of channels and channel networks from high-resolution digital elevation data is introduced in this paper. The proposed approach incorporates nonlinear diffusion for the preprocessing of the data, both to remove noise and to enhance features that are critical to the network extraction. Following this preprocessing, channels are defined as curves of minimal effort, or geodesics, where the effort is measured on the basis of fundamental geomorphological characteristics such as flow accumulation area and isoheight contours curvature. The merits of the proposed methodology, and especially the computational efficiency and accurate localization of the extracted channels, are demonstrated using light detection and ranging (lidar) data of the Skunk Creek, a tributary of the South Fork Eel River basin in northern California. |
2009 |
Roering, Joshua J; Stimely, Laura L; Mackey, Benjamin H; Schmidt, David A Using DInSAR, airborne LiDAR, and archival air photos to quantify landsliding and sediment transport Journal Article Geophysical Research Letters, 36 (19), 2009. Abstract | Links | BibTeX | Tags: DInSAR, landslide, LiDAR, sediment transport @article{Roering2009, title = {Using DInSAR, airborne LiDAR, and archival air photos to quantify landsliding and sediment transport}, author = {Joshua J. Roering and Laura L. Stimely and Benjamin H. Mackey and David A. Schmidt }, url = {https://angelo.berkeley.edu/wp-content/uploads/Roering_2009_GeophyResLet.pdf}, doi = {10.1029/2009GL040374}, year = {2009}, date = {2009-10-15}, journal = {Geophysical Research Letters}, volume = {36}, number = {19}, abstract = { We demonstrate the ability of coupled remote sensing tools to characterize large, slow-moving landslides in the Eel River catchment, northern California. From a stack of ALOS interferograms, we identified 5 large (>1 km long) landslides that exhibited significant activity from February 2007 to February 2008. For the Boulder Creek earthflow, we used orthorectified air photos taken in 1964 and unfiltered airborne LiDAR flown in 2006 to map the displacement of trees growing on the landslide surface. Combining those displacement orientations with stacked DInSAR data, we observed average downslope velocities of 0.65 m yr−1 through the central transport zone of the landslide. Given landslide depth estimates, minimum sediment transport and denudation rates are estimated to be 4100 m3 yr−1 and 1.6 mm yr−1, respectively. Our results demonstrate the highly erosive role of large, slow-moving landslides in landscape evolution and suggest that the superposition of dense, ephemeral gully networks and rapidly moving zones within the landslide may facilitate delivery of slide-mobilized sediment into adjacent fluvial channels.}, keywords = {DInSAR, landslide, LiDAR, sediment transport}, pubstate = {published}, tppubtype = {article} } We demonstrate the ability of coupled remote sensing tools to characterize large, slow-moving landslides in the Eel River catchment, northern California. From a stack of ALOS interferograms, we identified 5 large (>1 km long) landslides that exhibited significant activity from February 2007 to February 2008. For the Boulder Creek earthflow, we used orthorectified air photos taken in 1964 and unfiltered airborne LiDAR flown in 2006 to map the displacement of trees growing on the landslide surface. Combining those displacement orientations with stacked DInSAR data, we observed average downslope velocities of 0.65 m yr−1 through the central transport zone of the landslide. Given landslide depth estimates, minimum sediment transport and denudation rates are estimated to be 4100 m3 yr−1 and 1.6 mm yr−1, respectively. Our results demonstrate the highly erosive role of large, slow-moving landslides in landscape evolution and suggest that the superposition of dense, ephemeral gully networks and rapidly moving zones within the landslide may facilitate delivery of slide-mobilized sediment into adjacent fluvial channels. |