Dynamic Surfaces
and Geospatial Processes

Helena Mitasova

A. Petrasova,     V. Petras,     B. Harmon,     P. Tabrizian,     J. Jeziorska

   

An incomplete tour through evolution of surface modeling in open source GIS from its beginnings to current state and applications

The beginnings

Pioneering concept of dynamic spatial surfaces

Waldo Tobler:

geospatial phenomena as continuous scalar and vector fields

Images from Professor Tobler's slides

Theoretical foundations: terrain analysis

Terrain as continuous field z=f(x,y),
represented by isolines, manually deriving secondary fields

Profle Curvature isolines (Krcho 1973),

Computational foundations: digital terrain models

EuroCarto III Graz, 1984: Krcho,J., H.Mitasova, E.Micietova, Theoretical concept and data structures of complex digital terrain model and its interdisciplinary applications



Thank you for supporting our participation 34 years ago

Digital terrain modeling with splines

Regularized Spline with Tension (RST) interpolation: from scattered points to regular grids, with simultaneous derivation of topographic parameters


Profle Curvature draped over mesh surface, (Mitasova, Mitas, Hofierka 1993)

Implementation in open source GRASS GIS

Method became available to broad community, many developers improved the code

Surface with changing tension animation: tension helps to control overshoots

25 years: GRASS4.1 s.surf.tps, v.surf.tps, v.surf.rst; GRASS7.4; contributions of 8+ developers

Implementation in open source GRASS GIS

Quadtree-based segmentation made it applicable to large point data sets

Simultaneous topographic analysis: slope, aspect, curvatures

       

Modeling and analysis of surfaces from multitemporal observations

From digitized contours to lidar point clouds

Lidar technology transformed topography mapping in 21st century

Do we still need interpolation?

North Carolina lidar surveys

Coastal surveys started in 1996, second statewide survey just finished.
Survey in 1999 included Jockey's Ridge dunes: dense, noisy scattered point cloud

DEM: 1m resolution, binned

North Carolina lidar surveys

Coastal surveys started in 1996, second statewide survey just finished.
Survey in 1999 included Jockey's Ridge dunes: dense, noisy scattered point cloud

DEM: 3m resolution, binned

Regularized spline with tension for lidar

RST smoothing properties and quadtree segmentation made it suitable for generating high resolution elevation models from millions of points

DEM: 1m resolution, interpolated

Regularized spline with tension for lidar

By tuning the tension coupled with smoothing, we remove noise and derive topographic parameters at a desired level of detail:
profile curvature and slope draped over surface with changing tension

GRASS temporal framework for DEM time series

DEM time series for Jockey's Ridge dunes derived from data acquired by
  • photogrammetry (1974-98),
  • lidar (1999-2015),
  • structure from motion from UAS imagery (2016-2017)


Gebbert, S., Pebesma, E., 2014. A temporal GIS for field based environmental modeling. Environmental Modelling and Software 53, 1-12.

DEM time series visualization

Jockey's Ridge 1974 - 2017: southward migration, landform transformation from crescentic dune to sand starved, fast moving parabolic dune

DEM time series visualization 1997 - 2017

Contours time series

Contours capture the landform change but they are hard to read
16m
20m

Space-Time cube visualization

DEM time series is converted into space-time voxel model in TGRASS and evolution of a contour is represented as isosurface: 16m and 20m

Jockey's Ridge evolution analysis

DEM time series: evolution quantified using TGRASS and surface analysis tools

  • linear trend in loss of peak elevation at 0.3m per year
  • accelerating horizontal migration from 3m/yr to 6m/yr
  • total sand volume is stable, but the core (sand that has not moved) is shrinking
  • management challenges: dynamic feature confined to static park boundaries

From observations to modeling of processes

Water flow, soil erosion, storm surge

Land surface controls water and sediment flow across landscapes

Critical processes and impacts: surface runoff, flooding, storm surge, soil erosion

Water flow

Using surface gradients to compute flow accumulation: Evolution of water depth over complex terrain under steady rainfall and uniform surface conditions

Geometry-based solution

Sediment transport

Combining flow accumulation and slope:
Evolution of sediment transport capacity

Geometry-based solution

Erosion and deposition

Net erosion and deposition computed as change in sediment transport capacity:
simple to compute in GIS, combined with parameters for landcover and soils

Unit Stream Power Based Erosion-Deposition, Mitasova et al. 1995, Mitas and Mitasova 1998

Path sampling of continuous fields

Next step: robust solution of shallow water flow equations and process-based sediment transport

   

Solver based on duality of particles and fields works for noisy surfaces, captures ponding in depressions

Modeling surface runoff and erosion/deposition

Impact of construction on erosion and deposition, limitations of stream buffer protection

High resolution water flow

Street level modeling of surface runoff: lidar-based DEM and path sampling

Ultra-high resolution water flow: UAS mapping

Modeling impact of tillage and rills on surface water flow using 0.2m DEM derived by SfM based on UAS imagery

Updating lidar DSM using UAS based SfM

2015 lidar updated with 2018 UAS data: forested are replaced by a new school

Tangible interface for surface analysis and process modeling

Tangible Landscape

Bringing people together around GIS: Tangible user interface for GRASS GIS

Designed to make working with geospatial data and simulations engaging, and fun

Petrasova, A. et al. (2018). Tangible Modeling with Open Source GIS. Second edition. Springer International Publishing. https://doi.org/10.1007/978-3-319-89303-7

How does it work?

   

Tangible Landscape couples a digital and a physical model through a continuous cycle of 3D scanning, geospatial modeling, and projection

Interactions

Coupling with 3D rendering

Design scenario analysis

Tangible Landscape for designers and researchers

Tangible Landscape for education

Tangible Landscape for communities

Platform for decision-making and science communication where people of different backgrounds can interact.


Tangible Landscape website: tangible-landscape.github.io
TL wiki: github.com/tangible-landscape/grass-tangible-landscape/wiki

Students are making an impact

Integration of open source GIS in courses

Project-based graduate courses since 2008

Onset of bridge flooding

William Ross for NC disaster management

I-40 as a river in NCDOT drone video

County map of flooded bridges

William Ross for NC disaster management

Flooded bridges mapped based on lidar, bridge and inundation data for 100 yr storm

Improving storm surge flooding predictions

Nelson Tull for operational storm surge prediction using ADCIRC
Hurricane Matthew Hindcast - number of buildings flooded: before enhancement: 2435; after: 3886, a 60 percent increase.

N. Tull, J.C. Dietrich, T.E. Langan, H. Mitasova, B.O. Blanton, J.G. Fleming, R.A. Luettich 2018, Improving Accuracy of Real-Time Storm Surge Inundation Predictions Using GRASS GIS, Poster at ASBPA conference, October 2017. (American Shore and Beach Preservation Association)

Beyond bare earth surface

Individual tree detection from lidar

Geomorphons (Jasiewicz, Stepinski 2013) applied to vegetation surface: peaks represent individual trees. Detected trees are replaced by modeled trunks to improve accuracy of viewscape analysis (Tabrizian et al. 2018)

Beyond bare earth surface: vegetation voxel models

Generalized Fragmentation Index derived from 3D grid point counts: vertical slice of raw point cloud and slice of fragmentation index 3D raster

Beyond bare earth surface: vegetation voxel models

Slicing through fragmentation index 3D raster

Petras, V., D. J. Newcomb, and H. Mitasova. 2017. Generalized 3D fragmentation index derived from lidar point clouds. In: Open Geospatial Data, Software and Standards 2(9). DOI 10.1186/s40965-017-0021-8

Beyond bare earth surface: urban topography

Solar irradiation during summer solstice at NCSU Centennial Campus

Mapping pedestrian density from webcams

Georeferencing location of pedestrians from public webcams

Petrasova A., Hipp J. A., Mitasova H., Visualization and analysis of active transportation patterns derived from public webcams. FOSS4G 2017 - Boston, August 2017.

Space-time cube visualization of pedestrian density

Time series of predestrian locations is transformed into continuous space-time model using kernel density function

Selected density value visualized as isosurface in space-time cube

Change in pedestrian density (2014 minus 2012)

Positive values ~ increase in density in 2014, Negative values ~ decrease in density in 2014

Open Science

Developing open source software and contributing to OSGeo projects:

GRASS GIS https://grass.osgeo.org/
Tangible Landscape tangible-landscape.github.io

Open access educational material:

NCSU GeoForAll Lab Courses and Workshops https://geospatial.ncsu.edu/geoforall/courses.html

Thank You!

Thank you all for your contributions to the field - data, methods, algorithms and tools, that helped to bring the discipline to its current thriving state

Appendix

Links to related talks: TL webinar, GRASS7, Lidar, ICC talk etc.