Digital Terrain Analysis with LiDAR - Workshop Presentation · 2018-06-29 · 1 1 Digital Terrain Analysis with LiDAR for Clean Water Implementation May 11 2010May 11, 2010 Mankato

1

1

Digital Terrain Analysiswith LiDAR for

Clean Water Implementation

May 11 2010May 11, 2010Mankato

2

Digital Terrain Analysis with LiDARfor Clean Water Implementation

Workshop Introduction

Adam Birr, Ph.D.Minnesota Department of Agriculture

Research [email protected]

507-206-2881

3

Identifying Critical Portions of the Landscapefor Water Quality Protection

Using Terrain Analysis

Adam Birr, Ph.D.

Barbara Weisman

Minnesota Department of Agriculture

David Mulla, Ph.D., Jake Galzki, and Joel Nelson

Department of Soil, Water, and Climate

University of Minnesota

In accordance with the Americans with Disabilities Act, an alternative form of communication is available upon request. TDD: 1-800-627-3529 An Equal Opportunity Employer and Provider

2

4

Clean Water Fund

• Constitutional Amendment provides much-needed funding for protection, restoration, enhancement

• Clean Water Fund implementation dollars should be spent on the most critical landscapes and sources of water quality degradationwater quality degradation

• There is a pressing need for tools to identify these critical areas

5

Precision Conservation

• “…set of spatial technologies and procedures to implement conservation management practices that integrates spatial and temporal variability across natural and agricultural systems.” (Berry et al. 2003)

• “Getting the right practices, in the right places, at the right time, and at the right scale is what makes conservation effective.” (Cox 2005)

6

Critical Areas

• Studies suggest that small areas of the agricultural landscape (5-25%) generate a disproportionately large amount of erosion or phosphorus

Photograph Courtesy of the Brown, Nicollet, Cottonwood Water Quality Board

or phosphorus

• Defined as the portion of the landscape where the least amount of BMP investment yield the most benefit (Maas et al. 1985)


3

7

•• Two criteria:Two criteria:

–– Accumulation of Accumulation of surface flowsurface flow

–– Hydrologic Hydrologic connection toconnection to

Side Inlet

Gully

Critical Areas

connection to connection to surface waterssurface waters

y

8

Identifying Critical Areas• Challenging due to hydrologic complexity and

natural variability across landscapes

• Terrain analysis of Digital Elevation Models (DEMs) may be the key

• LiDAR data greatly enhances our ability to identify critical areas

9

• 25 Minnesota counties currently have LiDAR data

• Statewide LiDAR acquisition funded by Clean Water Fundby Clean Water Fund ($5.6 m)

• Acquisition to be completed in 2012

“The legislature got you the dollars, now go out and show us what you can do!”- Representative Rick Hanson


4

10

Digital Terrain Analysis Characteristics

• Sacrifices physical sophistication for simple calculations describing soil moisture patterns

• Input requirements are readily accessible and appropriate for the precision with which many management questions need to and can be answered (Barling et al 1994)questions need to and can be answered (Barling et al., 1994)

• Several studies have demonstrated the use of topographic indices to characterize the spatial distribution of soil moisture and soil mapping components controlled by soil hydrology (Bell et al., 1994; Thompson et al., 1998; Fried et al., 2000)

11

Clean Water Legacy Pilot Study

• Developed a tool that uses terrain attributes to identify critical source areas vulnerable to surface water runoff

– Focused on near-stream UPLAND features

– Piloted at multiple scales (3,000 ac to 800,000 ac)

using different DEM sources and resolutionsusing different DEM sources and resolutions

• 30m DEM for Le Sueur Watershed 8-digit HUC

• 3m LiDAR-derived DEM for Beauford Ditch Watershed (Blue Earth County) and Seven Mile Creek Watershed (Nicollet County)

12


Overview of Methods

• Calculated a suite of primary and secondary terrain attributes in the pilot watersheds

• Conducted a field survey to related terrain attributes to critical source features in the field

• Determined which terrain attributes are most applicable

• Used statistics to define threshold values for prioritizing critical areas


5

13Clean Water Legacy Pilot Study

Study Area2 pilot watersheds in Minnesota River Basin

14


Field Surveys

• Handheld Pocket PC with WAAS GPS

• Field mapping software

• Tape measure

• Digital camera

• Compass

• Log book

15Example: Using Flow Accumulation to Identify GulliesPilot Study Watershed: Seven Mile Creek (Nicollet County)


6

16Example: Using Flow Accumulation to Identify GulliesPilot Study Watershed: Seven Mile Creek (Nicollet County)

17

Courtesy of the Brown, Nicollet, CottonwoodWater Quality Board

18Example: Using Flow Accumulation

to Identify Critical Source Areas (Side Inlets, Gullies, Open Intakes)

Pilot Study Watershed:Beauford Ditch (Blue Earth County)


7

19Example: Using Flow Accumulation to Identify Critical Source Areas(Side Inlets, Gullies, Open Intakes)

Pilot Study Watershed: Beauford Ditch (Blue Earth County)

20

21


8

22Clean Water Legacy Pilot Study

Field Survey Results• Confirmed that ordinal size of a feature is related to Stream Power Index• Though not a quantitative assessment of sediment delivery potential, values

suggest a relationship between terrain attribute values and magnitude of erosional features

• Further study—including water quality data—needed to quantify relationship

SDP Score Average Percentile of SPI

Average Percentiles of SPI for Gullies in Seven Mile Creek Watershed, summarized by SDP Score

Side Inlet Size Average Percentile of SPILarge (24 - 36 inches) 98.9

Medium (14 - 18 inches) 93.3Small (4 - 12 inches) 81

gHigh (SDP = 3) 97.4

Moderate (SDP = 2) 83.8Low (SDP = 1) 72.8

Average Percentiles of SPI for Side Inlets in Beauford Ditch Watershed, summarized by Inlet Size

23


Conclusions

• Terrain analysis can be a very fast and effective way to identify critical areas

• Terrain attribute values indicate the ordinal, or relative, size of erosional features

– Can help us target conservation efforts to the most severe erosion risks

• Terrain analysis is easy to employ and a valuable use of newly acquired LiDAR data

24

Questions?

Adam Birr, Ph.D.Research Scientist

Minnesota Department of Agriculture507-206-2881

[email protected]


9

25


Workshop Lecture 1:LiDAR & Digital Elevation Models (DEMs)

Joel Nelson - instructor

26

Models of Topography:Elevation Data in General

• Stereo photography

• Topographic maps (elevation contours)

• Digital Elevation Models (DEMs)Digital Elevation Models (DEMs)– Stereophotograph models

– Ground survey (GPS, other)

– Interpolation from topographic maps

27

Stereo Photography

• View shape of topographic surface

• Overlapping photographs

• View from two perspectives V p p(parallax)

• Old technology – has been used extensively in Soil Survey and forestry


10

28

Topographic/Contour Data

• Also called contour maps

• Contour line joins points of equal elevation

• Can interpret slope, relief, shape/size of valleys and hills

• Paper and digital

– Digital leaves visualizationup to the user

29

Digital Elevation ModelWhat is a DEM?

• Digital file that:

• Contains elevation of terrain over a specified area

• Is arranged as a fixed-grid g ginterval over the earth surface

• Is geo-referenced

• Can be manipulated to create other elevation-dependentdata products

30

Digital Elevation ModelWaseca, Minnesota


11

31

Elevation

WasecaMinnesota

AgroecoResearch Farm,

100 m

HighLow

32

DEMSources

• USGS 7.5m series digitized contoursg

• Field survey

• Stereo photography

• GPS

• LiDAR (Light Detection and Ranging)

33

DEMCharacteristics

• ResolutionD it f l ti t– Density of elevation measurements

– Determines level of detail of surface representation

• Interpolation– Calculation used to find elevation of unspecified location

– Various techniques/algorithms: Kreiging, Theissen Polygons, Spline, IDW, Bilinear, Nearest Neighbor


12

34

Effect of Cell Size - Resolution

35

DEM Comparison

Why so much interest in LIDAR?

• Higher resolution data than we ever thought possible

• Opens up opportunities to describe and characterize landscapes in ways previously not feasible

Comparison to existing national standard product

USGS DEM LiDAR DEM

Horizontal Resolution 30 meters 1 meter

Vertical Resolution 7‐15 meters 15 cm

Contour Interval 5‐20 feet 1‐3 feet

36

DEM Comparison

USGS 30m DEM LiDAR 3m DEM


13

37

DEM Comparison

38

USGS 30 meter Elevation Data

39

LiDAR 3 meter Elevation Data


14

40

41

Contour Comparison

2 ft contours created from LIDAR data

10 ft contours createdfrom standard 30m DEM data

42

LiDAR

What is LiDAR?

• Light Detection And Ranging – a remote sensing system used to collect topographic datap g p

• Used by agenciesand organizations to develop high-resolution elevation information


15

43

LiDAR Resolution

• Bare earth resolutions– Vertical: 15cm

– Horizontal: 1m• Root Mean Square Error

(Reutebach et al 2005)(Reutebach, et al., 2005)

44

LiDAR Data Collection

• Aircraft mounted laser(s)

• LiDAR sensor records time difference between laser

How is LiDAR data collected?

difference between laser beam emission and return of reflected laser signal to aircraft

• Each pulse has an associatedreal-world GPS locationwww.esri.com/library/userconf/proc00/professional/papers/Pap808/p808.htm

45

How is LiDAR data collected?• End-product is accurate,

with geographically registered longitude, latitude, and elevation (x y z) for every data point

LiDAR Data Collection

(x,y,z) for every data point

• LiDAR data collection technology is rapidlyimproving

• Several file types and derivative productsavailable to end-users

www.esri.com/library/userconf/proc00/professional/papers/Pap808/p808.htm


16

46

LiDAR Data CollectionLiDAR Returns: Multiple discrete return pulses

LiDAR Intensity: Magnitude or strength-of-return pulse

Metadata: Information about how data was collected—READ IT!

• All returns can be used• Forest canopy

I i i• Intensity image• Vegetation mapping

47

DEM Pre-Processing• Techniques/tools to prepare dataset for analysis

• LiDAR data is changing this step

• Pre-processing tools

– Pit Filling

Elevation

High

Low

– Filter

– HydrologicConditioning

48

Pit Filling• Artificially draws base elevation levels in “sinks”

or “peaks” to bank-height or surrounding elevation values

• Usefulness / appropriateness depends on landscape and datalandscape and data


17

49

Pit Filling

• Useful for:– Removing anomalies and erroneous values– Closed-depression landscapes– Flood drainage scenarios – fills up depressions,

forcing water to flow over

• Caveats:– Data and drainage is being

altered, made artificial– Depicts accurate flow only

at flood stages or higher

50

Pit Filling

51

Filter Analysis

• Low Pass Filter – a moving-window analysis that performs neighborhood averaging (9-cell basis)

• Low-pass smoothes extremes, High-pass enhances them


18

52

Filter Analysis• Useful for:

– Improving DEMs by eliminating spurious data– Smoothing out local variations, providing more

connectivity to results

• Caveats:– Averages out data extremes– “Dumbs down” the data– Results in “fuzzy effect” –

resolution appears reduced even though it is not

53

Hydrologic Conditioning• Any alteration that improves flow-thru/drainage

• Several types, from manual to fairly automated

• Common types:– Stream burning—uses stream data to force drainage

– Bridge/obstruction removal involves manual editing– Bridge/obstruction removal—involves manual editing,but becoming more automated

– Hydrologic breaklines—optional/additional data used toidentify sharp changes in relief

54

Hydrologic Conditioning• Useful for:

– Routing flow through an entire watershed ordrainage network

– Correcting erroneous river channel boundaries

• Caveats:– Stream data/other data used with

LiDAR must be same resolutionor spatial quality

– Flow/drainage is often site-specific and dependent upon the quality of the control structure


19

55

End of Lecture 1

Questions?

56


Lecture 2:Terrain Analysis


57

Terrain AnalysisWhat is it? Many things:• Includes use of a DEM to model

the landscape

• Provides a quantitative, detailed, objective, repeatable process to accurately model real landscape processeslandscape processes

• Coupled with GIS/Remote Sensing technologies, allows us to quickly and accurately characterize large areas

• Helps describe, analyze, and interpret any feature related to topography – soils, vegetation, wildlife, etc.


20

58

Terrain Analysis History

• Concept is over 20 years old

• Early pioneers Wilson, Gallant, Moore, Gessler

– Terrain Analysis: Principles andApplications (see references, last slide)Applications (see references, last slide)

• Early applications:– Soil Mapping

– Hydrologic Mapping

– Wildlife/Habitat Modeling

• LiDAR is creating renewed interest in terrain analysis

59

Advantages of Terrain Analysis

• Coupled with GIS/remote sensing, enables fast, accurate characterization of large areas (days vs. months)

• Quantitative, repeatable, and non-subjective

• Helps describe, analyze, and interpret any featurerelated to topography (soils, vegetation, wildlife, etc.)

• Results in spatial data, not just numerical data

• Fits the level of detail needed for conservation planning

60

Cost Benefits of Terrain AnalysisPilot Study Watershed – Seven Mile Creek

• Walking survey: 10 days, ~300 hours, 3 people

– Total cost was $9,500, or ~$413/ditch mile

• To field-survey the rest of Nicollet County would take about 10 years and ~$100,000 in labor

Source: Brown Nicollet Cottonwood Water Quality Board


21

61

Cost Benefits of Terrain AnalysisPilot Study Watershed – Seven Mile Creek

• LiDAR-based GIS survey was completed ina matter of hours

• A county-wide survey identifying most of the largest contributing areas could be done in a matter of weeks

• Terrain analysis requires far less time and resources than field-based surveys

62

Terrain Attributes• Primary and secondary

• Primary attributes calculated directly from elevation data– Examples: Aspect, Slope, Flow Accumulation,

Profile Curvature, Plan Curvature

• Secondary (compound) attributes involve combinations of primary attributes – the are indices– Indices describing the spatial variability of specific landscape

processes, such as the potential for sheet erosion (Moore et al., 1991)

– Examples: Stream Power Index, Wetness Index

63

Digital Terrain Analysis Overview Example

DEM

Terrain

Point Elevationswith GPS/LiDAR

TerrainAttributes

SpatialInterpolation

AttributeCalculation


22

64

Primary Terrain Attributes

Aspect• Describes the cardinal direction a surface

faces (0°-360°)

• Uses:

– Fire management

– Soil moisture andevapo-transpiration

– Flora and fauna distribution and abundance

65


Slope

• Describes overlandand subsurface flowvelocity and runoff raterate

• Quantifies maximum rate of change in value from each cell to its neighbors

66

Slope

Blue Earth CountyMi t

Beauford DitchSubwatershed,

HighLow

Minnesota


23

67

Slope

– Overland and subsurface flow

– Velocity and runoff rate

– Precipitation

• Use/Significance

p

– Vegetation

– Geomorphology

– Soil water content

– Land capability class

68Primary Terrain Attributes

Curvature

Plan Curvature• Measured perpendicular to

direction of descent

• Describes convergingor diverging flow

Profile Curvature• Measured in direction of

maximum descent or aspect direction

• A measure of flow acceleration, i /d iti t• Contour curvature erosion/deposition rate

69

Plan

Curvature

Profile

Convex

Concave


24

70

Curvature

Use/Significance

• Plan Curvature– Converging or

diverging flow

– Soil water contentSoil water content

– Soil characteristics

• Profile Curvature– Flow acceleration

– Erosion/deposition rate

– Geomorphology

71


Flow Accumulation• A measure of surface or

shallow subsurface runoff at any given point on the landscape

• Combines effects of upslope surface drainage area and convergence of runoff

• Also called Catchment Area

72


Flow Accumulation

• Represents drainage area of any given cell

• Indicates overland flow paths

• Also called Catchment Area or Upslope Contributing Area

Elevation

300 m

308 m

30

10,000

Flow Accumulation


25

73

Flow Accumulation

Blue Earth

Beauford Ditch Subwatershed,

CountyMinnesota

HighLow

74

75


Flow Accumulation

– Runoff volume

– Steady-state runoff rate

Use/Significance

y

– Soil characteristics

– Soil-water content

– Geomorphology


26

76

Secondary Terrain Attributes

• Second derivative calculations(combinations of primary terrain attributes)

C d T hi I d (CTI)Compound Topographic Index (CTI)

Stream Power Index (SPI)

77


Stream Power IndexSPI = ln (A * Slope)

Compound Topographic IndexCompound Topographic IndexCTI = ln (A / Slope)

where A = Flow Accumulation

78


Stream Power Index (SPI)

• Product of Slope and Flow Accumulation

• Quantifies the potential erosive power of overland flow

• Isolates areas with large catchments and steep slopes

ln (A * Slope) = Stream Power Index (SPI) SPIln (A Slope) Stream Power Index (SPI)

X =

High

Low


27

79

Stream Power Index

• Measure of potentialerosive power of overland flow

• Combines Flow Accumulation (catchment area) with slope

• Steep slope with large drainage area results in high SPI value

– Indicator of where ephemeral gullies may form in a field

80


Compound Topographic Index (CTI)

• Flow Accumulation divided by Slope

• Identifies areas where water collects or ponds on the landscape

• Also called Steady State Wetness Index or just Wetness Index

7

20

=

ln (A / Slope) = Compound Topographic Index (CTI)

81

Upland DepressionsCritical Areas

Compound Topographic Index

• Measure of potential wetness in the landscape

• Combines Flow Accumulation (catchment area) with slope

High CTI Value

• Low slope and/or high catchment equal high potential for water topond or collect– Indicator of potential

wetlands and varioussoil types


28

82

End of Lecture 2

Questions?

83


Lecture 3:Performing Terrain Analysis


84

Performing Terrain Analysis – Principles

• Dependable method to quantify anddescribe the landscape

• Iterative process that builds on what’slearned in previous steps

• Adaptable – dependent on landscape feature of interest

R l i i ib i di• Relative – terrain attributes are indices

• Dynamic – improvements to algorithms and software enable solutions to previous problems and issues– There’s an overall process, but individual detailed steps

not necessarily the same each time

– Should always be supplemented with field visits, ancillary data, or site-specific knowledge


29

85

Process

Planning/Goals

Pre-Processing

No

Calculate Primary Ancillary Data

Yes

Calculate Primary

• Pit Fill• Filter• Stream Burn• Other

Terrain Attributes

Calculate Secondary Terrain Attributes

Ancillary Data

Ground Truth

Comparison

Prioritizing

Visualize/Report

DECISION

Terrain Attributes

Calculate Secondary Terrain Attributes

86

Pre-Planning Strategies

Analysis goals? Develop these first.

• Pre-processing steps and progression of calculations may differ depending on:

– Landscape p

– End products – e.g., flow network of watershed vs. specific erosion problem areas

– Focus – e.g., uplands, lowlands, or both

– In-house vs. public audience – presentation of terrain attributes can be complex

87

Pre-Planning Strategies

• Landscape – physiographic characteristics of thelandscape modeled willinfluence decisions

• High-relief, dendriticnatural drainage– Flow networks are

fairly contiguous

– Pit filling may be appropriate(few natural pits in this landscape)


30

88

Pre-Planning Strategies• Landscape – physiographic

characteristics of thelandscape modeled willinfluence decisions

• Low relief,depression-filled landscape– Flow networks often

irregular local

Restorable Wetland Polygons

Upland Depressions Critical Areas

irregular, local– Pit filling, stream

burning can mimicflood stage flow, nottypical conditions

– Run both filled and unfilled

Data source: Ducks Unlimited, Inc. MN DNR, & MN LMIC

89

Ancillary Data

• Terrain attributes are based solely on topography

• Use other information to aid decision-making– Land use data

– Land management data

A i l h t h– Aerial photography

– Existing conservation

practices

– Distance to water

90

Ancillary Data

Aerial Photography• Particularly useful

• In limited situations, can useto validate/verify groundconditions

• Higher quality orthophotosf fl i h i doften flown with LiDAR data

• Can also clarify unexplainedanomalies or uneven terrain

• Does not replace ground-truthing

• Dated product


31

91

Ground Truthing• Very necessary to relate mapping to planning goals• Important step in comparing terrain attribute threshold values to real-

world conditions• Strategies

B i d hi– Basic ground truthingoverall

– Randomly select areas to do exhaustive surveys

• Windshield surveys

• Detailed surveys

92

Comparison

• Identify points where terrain attributes indicate area of interest

• Compare attribute

GIS based survey points

Stream

SPI ≥ 85th percentile

Compare attribute values with ground truth data at same location

93

Comparison• Compare real-world features to those identified by

terrain analysis (e.g., gullies, depressions, etc.)

• Create table/matrix to record errors• Type I Error / False Positive – terrain analysis indicated a

feature that wasn’t there

• Type II Error / False Negative – terrain analysis indicated no feature, but one existed


32

94Prioritizing

Percentile Ranking• Terrain attributes converted to percentiles and ranked

• Allows comparison among all values within a given area (watershed, county, etc.)

• Comparison is more uniform than guessing random values

• Visually depicts spatial distribution of upper X% of specific terrain attribute valuesterrain attribute values

95

Prioritizing

Percentile Ranking

• Plot of all SPI values in watershed

• Side Inlet locationsSide Inlet locations correspond to upper-endSPI values

96Prioritizing

Percentile Ranking• How-to • First extract all individual cell records – export point• Several methods

– Statistical Software Package– Excel 2007 Percentile command – very simple

• Limited to a little over 1 million records• LiDAR-derived terrain attributes often have more than 1 million records


33

97

Prioritizing• Big-picture analysis to account for all that

terrain analysis doesn’t– Proximity to water– Landscape position relative to ancillary data of interest– Location within the watershed– Costs: Survey, Remediation/Attenuation, Labor

98

Visualization• GIS – Terrain Analysis

• Identify via combination of attributesand your own criteria– Alter symbology for attributes

– Display upper X% of valuesthat satisfy goals/needs

0 500 1,000250Meters

0 5 102.5Kilometers

Ravine Critical Areas

Minnesota River

Minnesota River Tributaries

99

Terrain Analysis Caveats/Limitations

• Same limitations as LiDAR data in general– Cost– File Size/Computing Power– Expertise/Training

• Terrain analysis does NOTTerrain analysis does NOT……Replace local knowledge…Make fieldwork unnecessary or obsolete…Transfer well to non-like landscapes when comparing terrain

attribute values…Differentiate between man-made and “natural” structures


34

100

Terrain Analysis Caveats/Limitations

• Pertains to surface flow only

• No automated or one-size-fits-all process

• Results only as good as what you put into itD– Data

– Assumptions

– Effort

– Ground truthing / Error checking

101

End of Lecture 3

Questions?

102

Credits/Acknowledgements• Minnesota Dept. Agriculture

– Dr. Adam Birr– Barbara Weismann– Mike Dolbow– Jim Gonsoski– Karl Hillstrom– Brian WilliamsBrian Williams

• Jake Galzki – U of MN

• Dr. David Mulla – U of MN

• Dr. Jay Bell – U of MN

• Brown/Nicollet Cottonwood Water Quality Board

• Blue Earth County – Information Technology Department

• Goodhue County Survey/GIS Department


35

103

References• Gallant, J.C. and J.P. Wilson. 2000. Primary topographic attributes.

p. 51 - 85. In J.P. Wilson and J.C. Gallant (eds.) Terrain analysis: Principles and applications. John Wiley and Sons, Inc., New York.

• Gessler, P.E., I.D. Moore, N.J. McKenzie, and P.J. Ryan. 1995.Soil-landscape modeling and spatial prediction of soil attributes. International Journal of GIS. Vol 9, No 4, 421-432.

• Moore, I.D., P.E. Gessler, G.A. Nielsen, G.A. Peterson. 1993. Soil attribute prediction using terrain analysis. Soil Sci. Soc. Am. J. 57:443-452452.

• Moore, I.D., R.B. Grayson, and A.R. Ladson. 1991. Digital terrain modeling: A review of hydrological, geomorphological, and biological applications. Hydrol. Processes. 5:3-30.

• Tomer, M.D., D.E. James, and T.M. Isenhart. 2003. Optimizing the placement of riparian practices in a watershed using terrain analysis. J. Soil Water Conserv. 58:198-206.

• Wilson, J.P. and J.C. Gallant. 2000. Secondary topographic attributes. p. 87-132. In J.P. Wilson and J.C. Gallant (eds.) Terrain analysis: Principles and applications. John Wiley and Sons, Inc., New York.


Documents

Digital Terrain Analysis with LiDAR - Workshop Presentation · 2018-06-29 · 1 1 Digital Terrain Analysis with LiDAR for Clean Water Implementation May 11 2010May 11, 2010 Mankato