When:

-we combine data from different sources

-we present image data on an established map projection

Types?

-Raster image scaling (Moire effect - aliasing)

-Mosaicking after rectification

-Image binarization (making two shades, 0 and 1)

-Image Segmentation (ie. segment an air photo into polygons)

-Vectorization of scanned paper maps (similar to segmentation)... creates binary image then digitized curves then vector data

Is the science of obtaining reliable measurements from photography

The results can be:

-coordinates of the required object-points

-topographical and thematical maps

-and rectified photographcs (orthophoto)

national mapping strategy working group, Canadian council on geomatics (2010)

1. Growing realization that geography/geographic information is required in a modern world - needs of citizens

2. Long-term approach to planning and coordination.

3. National Standard for Spatial Data Accuracy established by teh Federal Geographic Data Committee in 1998

Importance of the National Strategy...

-Began in 2007- a little late by global standards

-GIS is the driving forec

-Internet and development of geographic data sharing

-Relationships between different levels of gov

-Growth of mass market applications presented policy and operational challenges for mapping organizations.

CONSULTATIVE AND INCLUSIVE GOVERNANCE

-Working cooperatively will provide clarity, direction, and certainty for Canada's mapping community by providing a forum to share information, methodologies, approaches to mapping and timelines.

NATIONAL APPROACH, REGIONAL AND LOCAL DECISION MAKING

-Effective and responsive policies that respect the autonomy of mapping agencies and organization while promoting cooperation and coordination are crucial to the imlpementation of the National Mapping Strategy

COMMON TECHNOLOGICAL FOUNDATION

-A major challenge for the National Mapping Strategy will be to ensure that the broad technical and data standards required for interoperability continue to be promoted and their adoption encouraged.

CURRENT AND AVAILABLE DATA

-New approaches to data maintenance and updating cycles will be designed and implemented to ensure that data sets required by decision makers are as current and accurate as required.

GEOGRAPHICAL DATA AS A PUBLIC ASSET

-The NMP recognizes the importance and value of 'no charge' unrestricted access to government mapping data when tehre is no threat created by the release of the data.

OUTREACH AND COMMUNICATIONS

-A concerted effort will be made to inform non-traditional users of Geomatics technologies, and to inform policy and decision makers of the enriched information and analytical power that can be obtained through the use of geomatics and how this can substantially benefit the work they do.

A VIBRANT GEOMATICS INDUSTRY

-A central componenet o the NMS is a competitive, productive geomatics industry, able to provide made0in Canada solutions, strongly benefiting the Can economy.

AN AVAILABLE AND EDUCATED LABOUR FORCE

-Promoting opportunities for education in geomatics, ranging from technologists, to engineers to analysts, in order that there are qualified Canadian workers to meet the challenges of the future is important for the sustainability of the community.

3289 companies reported as being involved in surveying and mapping in 2007 Stats Can survey

2.8 billion anual revenue

23000 Canadians employed by mapping sector

Location Errors

-refer to the geometric inacuracies of digitized features

-can be examined by referrnig to the data source for digitizing

Topological Errors

-violate the topological relationships either required by a GIS package or defined by the user

-topological erors with geometric features: undershoot, overshoort, dangling node (results from over/undershoot), pseudo node, direction error, label error

- a pseudonode is a node that is not located at a line intersection

-topological errors between layers: boundaries not coincident, lines not connected at end points, overlapping line features, etc.

Topological editing ensures that digitized spatial features follow the topological relationships that are either built into a data model or specified by the user.

1. Topological editing on coverages

2. Editing using map topology

3. Editing using topology rules

-Specified Dangle Length (can remove an overshoot)

-Special fuzzy tolerance (snap duplicate lines if the gap between the nlines is smaller than the specified tolerance)

-Allowable extend distance (close undershoots)

If thresholds too high, all of these can ruin data.

Refers to a variety of basic editing operations that modify simple features and create new features from existing features.

-Can reshape lines by moving, deleting, or adding vertices

-can split or merge polygons

-Edgematching (self explanatory)

-Line smiplification (process of simplifying by removing points)

-Line smoothing (reshaping lines by using some mathematical functions such as splines)

David Douglas and Tom Poiker 1973

-Given a curve composed of line segments, find a curve with fewer points

-Dissimilar based on maximum distance between the originalcurve and the simplified curve

-It is a recursive algorithm

DP Algorithm

-Marks the first and last point and keeps them

-Finds the point that is furthest from the line segment with the first and last points defining a line

-Compares the points to the distance dimension (tolerance of line variation)

-If the line meets the distance dimension criterion, it is kept... if not then points that exceed the distance dimension are kept.

-Iterates until all points in the new line meet the distance dimension

Leaves a sharp angle! Retains variability!

Blend and simplify is another line simplification algorithm.

Dissects the line into a series of bends, calculates teh geometric properties of each, and removes "insignificant ones"

Smoother results, but slower. Considers overall shape of the line

Result of line simplification can differ depending on the algorithm (ie. Douglas-Peucker vs bend-simplify)

Metadata is the contextual component of the data set.

It is data 'reporting'

-WHO created the data?

-WHAT is teh content of the data?

-WHEN was it created?

WHERE is it geographically?

HOW was the data developed?

WHY was the data developed?

ie. title, supplemetnal information, abstract, time period, author,  sources, file size, etc

FORMAT

-used to be text and html, now xml and xsl

Metadata is the second abstraction

-go from real world to map/image/db (abstraction 1- content), then go from content to context (abstraction two)

Metadata: Building Block of SDI

-Spatial Data Infrastructure: a set of actions and new ways of accessign, sharing, and using geographic data that enables far more comprehensive analysis of data to help decision makers choose the best course of action

-it allows us to access, share, and use context (metadata)

Refering to Metadata

-While all this can be imporatnt inforamtion, the GIS analyst typically needs to rely most on Major SSections 4 and 5

---Spatial Reference Information

---Entity and Attribute Information

-In ArcGIS, metadata is fairly closely tied to the Geospatial data it describes

-However, a brief version is shown by default and it is known as the Description

(ie. interpolation)

IDW

-measures distance between points, assignes a weight based on their distance

Parameters change results of IDW interpolation.

-The grid size

-the choices of neighbouring control points (how many neighbours, or circle radius)

-the distance weighting or distance squares

-The form of distance functions (ig. inverse negative exponential)

Spline: Finds contours that are the smoothest possible curves that can be fitted and still honor all the data

Kriging: Geostatistical method

-estimation of spatial assocaiteion (variogram)

-estimation of the point value using spatial association information

How to make the weights more sensible?

Kriging assumes taht the spatial variation of an attribute may consist of three components:

-a spatially correlated component

-a 'drift' or structure, representing a trend

-a random error term

ORDINARY KRIGING

-assumes the absence of a drift and focuses onthe spatially correlated component.

-the measure of the degree of spatial correlation or dependence among the sampled points is the average semivariance

-semivariance increases with distance

There are 5 mathematical modesl for fitting variograms: gaussian, linear, spherical, circular, and exponential.

USE of a Fitted Semivariogram

-ordinary kriging uses a fitted semivariogram directly in spatial interpolation.

UNVERSAL KRIGING

-assumes that the spatial variation in z values has a drift, or a trend, in addition to the spatial correlation b/w the sampled points

-it typically incorporates a first-order (plance surface) or a second-order (quadratic surface) polynomial in the kriging process.

(as opposed to numerical)

Different approach to surfacing

-can be global or local

-fit a mathematical surface to the set of points

-creates an equation of the surface

-equation can then be evaluated at any X,Y coordinate to yield a surface

Local Interpolators (window-based)

-Generates estimates based on existing data in the "region"

-"region" = "roving window"

--moves about study area

--summarizes data it encounters

--reach (search radius)

--number of samples

--direction

Minimum Curvature Splines

-fits a minimum-curvature surface through input points

-like bending a thin sheet of metal to pass through points, while minimizing curvature of that sheet

-repeatedly applies a smoothing equation (piecewise polynomial) to the surface, resulting surface passes through all points)

-Best for gently varying surfaces, not for rugged ones (can overshoot data values)

Thin-plate Splines (and with tension)

-A major problem with thin-plate splines is the steep gradients in data-poor areas, called overshoots.

-Thin-plate splines with tension (parameter) allows the user to control the tension to be pulled on the edges of the surface

-The larger the tension parameter, the smaller the stiffness of the plate (surface) and consequently the range of interpolated value

SPLINES cont...

The Regularized option yields a smoother surface

-the weight parameter defines the "weight of the third derivatives of the surface in the curvature minimization expression"

-the high its value the smoother the surface

The Tension option tunes the stiffnesso fthe surface according to the character of the modeled phenomenon

-the weight parameter defines the "weight of the tension"

-teh higher its value the coarser the surface

The number of points parameter identifies the number of points per region used for local approximation.

Higher values produce better surfaces (still exact fitting)

**Spline methods are best for gently varying surfaces such as elevation, water table heights, or opllution concentrations.

Not appropriate if there are large changes in the surface within a short horizontal distance, because it can overshoot estimated values.

Like IDW, it is an exact interpolation

UNLIKE IDW the surafce can extend above or below measured points

SPLINES in ArcGIS

Spatial Analyst offers Regularized Spline and Spline with Tension

Geostatistical Analyst offers Radial Basis Functions, which include Completely Regularized Spline, Spline with Tension, Multiquadric, Inverse Multiquadric, and Thin Plate Spline

RBF/Splines should be used for calculating smooth surfaces from a large number of data points

Comparison of local methods is usually based on statistical measures

Visual quality of generated surfaces such as preservation of distinct spatial pattern and aesthetics and faithfulness

Cross validation and validation are two common statistical techniques for comparison

CREATING A SURFACE OF ERROR

CROSS VALIDATION

Compares the interpolation methods by repeating the following procedure for each interoplation method to be compared:

1. remove a known pnt from dataset

2. Use the remaining points to estimate the value at the pnt removed

3. Calculate the predicted error of the estimate by comparing the estimated with known

VALIDATION

Comapres the interpolation methods by first dividing known points into two samples:

-one sample for developing the models for each interpolation method to be compared

-the other sample for testing the accuracy of the models

The diagnostic statistics of RMSE and the standardized RMSE derived from the test sample can then be used to compare the methods

Primarily created by photogrammetric methods from stereoscopic photography

Gov of Can - Nat Res Can - Centre for Teopographic Info are responsible for data compilation

DEM dl. from GeoBase

To visualize?

-change pallet

-addhillshade

-add colour

-change sun position

-etc

(chapter 14)

Involves combining terrain attributes to derive maps of important features

Involves spatial modelling

Widely used in Hydrological, Geomorphological, and Biological applications

General Approach to Modelling

-Not dealing with right - just a model

-Must have some domain knowledge of the feature being studied

-Data variables will control what can /can't be done

-simplest models are those that are controlled by terrain attributes

TERRAIN ANALYSIS

Attributes that can be computed from terrain:

-derivatives (slope, curvature)

-Watershed analysis (local and global catchment areas, stream delineation, catchment area)

-Solar radiation

-Visibility (viewshedS)

Overall goal is to compute topographic attributes to describe:

-morphometry (shape)

-catchment position

-surface attributes (erosion risk, wetness index)

-channel position

-pattern-process relationships

Surface Runoff is a function of upslope contributing area, soil transmisibity, and slope gradient.

If this statement is true, the bottom of a hillslope would be wetter than the top.

Generating the inputs

-generate contributing areas

-calculate for every cell in DEM

-based on an accumulation of values from a unity grid (all ones) and aspect

-use 8-direction pour point method

-each raster has a value, take direction of steepest descent

-end up with a flow direction grid

Flow Direction Grid

-Water flows to one of tis neighbour cells according to the direction of steepest descent

-Flow direction takes one out either possible values

-end up with a dendritic stream network, which allows for watershed delineation

RASTER vs VECTOR STREAMS

Some mis-match between digital stream networks and hydro lines

Procedures available to incorporate known vector locations of streams (i.e. Stream Burning)

"FILLING PITS" in the DEM

DEM creation results in pits (natureal and artificial) in the landscape.

A pit is a set of one or more cells which has no downstream cells around it

Unfilled pits become sinks and isolate portions of the watershed

Pit filling should be the first procedure performed on a DEM

THE VARIABLES

DEM and Derivatives

-pit filled DEM

-flow accumulation grid

-flow direction grid

-slope

-hillshade

TOPOGRAPHIC WETNESS

Assumes water runs down a hill

BAsed on flow accumulation grid and slope

Flow accumulation indicated up-slope contributing area

Slope yields the intensity/velocity

Should have soil transmissivity to modulate slope

STREAM POWER

Measure of erosive force of flowign water based on the following assumptions:

-Discharge is proportional to specific catchment area (up-slope contributing area)

-Flow erosive power will increase with slope

-All flow is overland (no base flow)

Predicts net erosion in areas of profile convexity and tangential concavity (flow acceleration zones)

Net deposition occurs in areas of profile concavity (decreasing flow velocity)

Modeling the amount of irradiance received at a point.

Itterative procedure based on sun angles and topographic surface interactions.

Modeled for single days, months, years.

Does not account for diffuse components - assumes perfect sky conditions.

SHould becorrected for yearly and monthly variations in irradiance.

Analytical Hill Shading

Traditional cartographic technique used to show topography

Uses slope and aspect

Assigns a soalr elevation and azimuth

-vary contrast by varying solar elevation and azimuth

Either Lambertian or Ray Tracing

Lambertian Relative Radiance

Computes the relative radiance reflected from a surface - use as an analogue for computing irradiance received by the surface

Similar in concepts to a watershed, but it contains what is visible.

Often called intervisibility

Based on line-of-sight analysis

How would we construct a theoretical model to eprform this analysis?

Intervisibility

Depends on:

-DEM characteristics

-DEM accuracy and errors

-Vegetation and buildings

-Atmospheric effects

-Geometric model

High resolution, readily available DEMs for intervisibility exist.

Required for Acceptable Algorithm

-UTM or lat/long DEMs

-variable observer height

-variable target height

-Understandable across communities (terrain analysts, GIS, modellers)

-publicly documented

Point to Poitn Viewsheds are the current "norm

-from the viewer's lcoation, run intervisibility to every grid point in the DEM

-Starts at farthest point and works back to start

EXTENSIONS TO VIEWSHEDS

Can add "target height"

It target is above the terrain it would be visible above teh normal terrain height

Allows for more meaningful interpretation (both viewer and target have a height)

Military uses are obvious

OTHER USES OF VIEWSHED

Urban Modelling

-visual aesthetics

-noise

-cellular and wireless networks

Forestry

-cut block planning

A GIS technique for evaluating the potential pathway of some feature.

Path analysis is raster based and involves measuring the "cost" associated with movement.

Often used as a planning tool to determine routings (roads, pipelines etc)

PATH MECHANICS

You need a starting point and an end point.

-some programs require separate grid inputs

Some measurement of "cost"

An algorithm for accumulating cost as it traverses from start to end

COST SURFACE

-simple distance

-topography

-geology

-stream/road crossings

-right of way

-population

-cultural resources

-land use

-vegetation and wildlife

Make a grid that has a total cost for each cell

Some factors could be attractors, some deflectors

Could create the alues based on reality, or index (difficult to put a value on all things)

MOVING THROUGH THE COST SURFACE

Physical Distance

Cost distance measures the cost of traversing the physical distance

-the cost can be determined by travel time, speed limit, congestion, terrain characteristics, vegetative cover, etc.

Distance measures in a grid follow the node-link relationship

A node representst he centre of acell, and a link can be either lateral or diagonal.

Distances are calculated along the links in a cell, where a lateral link is 1.0 cell (in cell resolution units) and a diagonal link is 1.41 cells.

There are two types of physical distance measure operations:

1)computation of linear distance from a source

2) allocation of area to its closest source

Source grid is the only input needed for physical distance measure operations

Straight line distance (euclidean distance) can be used to create buffers around cells or zones in a grid. Wave model.

Allocation grid determines for each cell in a grid its closest sources by the distance measure

MEASURES OF DIRECTION IN A GRID

Direction measures follow the eight principal directions from a focal cell to its eight neighbouring cells.

Expressed in degrees they are 0, 45, 90, 135, 180, 225, 270, 315

Same procedure used for flow modelling

Cost Distance Measure Operations (path analysis)

Cost distance measure operation requires two input grids:

1)a source grid

2)a cost (impedance) grid

Given a cost grid, the cost distance of a lateral link is the average of the costs in the linked cells

The cost distance of a diagonal link is the average cost times 1.4

***Why compute cost distance?

The objective of cost distance measure is to find the path with the least accumulative cost between a source and a destination

(wave-front algorithm, Dijkstra 1959, Tomlin 1990)

Activate cells adjacent to the source cell and compute costs to the source cell.

Select the cell with the lowest cost-distance to the source and assign its value to the output grid.

Activate cells adjacent to the selected cell, update the costs to the source cell, and again select the cell with the lowest cost.

Iterate through steps 1-2 until each cell has its lowest accumulative cost assigned.

Least accumulative cost path can only be derived after each possible path has been evaluated. Computing all possible paths is an iterative process.

1) Least accumulative cost grid

2) Direction grid, showing the least cost path from each cell to a source

3) Allocation grid, showing the assignment of each cell to a source on the basis of cost distanec measure

4) Shortest path grid, showing the least cost path from each destination cell to a source

Check it out! Could be question like this.

-determine criteria

-create a model flowchart

Step 1: Discrete preference Map

-identifies teh relative preference of locating a transmission line at any location throughout a project area considering multiple criteria

Step 2: Accumulated Preferences Map

-identifies the preference to construct the preferred transmission line from a starting location to everywhere in a project area

Wave algorithm - like tossing a stick into a pond with waves emanating out and accumulating costs as the wave front moves.

Step 3 - Most preferred Rout

-teh steepest downhill path over teh accumulated preference surface identifies the most preferred route - minimizes areas to avoid.

Calibrating Map Layers (Relative Preferences)

-model calibration refers to establishing a consistent scale from 1 (most preferred) to 9 (least preferred) for rating each map layer.

Generate alternative routes (always give 3 options)

-simple average

-community weighted

-environment weighted

AHP is used to establish relative importance among location criteria based on group values.

The procedure involves mathematically summarizing paired comparisons of the map layers' importance

Netowork: a set of interconnected line entities, whose attributes share some common theme primarily related to flow.

Network in a GIS: a topology-based line coverage

Geometric Network in ArcGIS geodatabase: a feature data set comprised of feature classes: edges and junctions, and their connectivity defined by topology rules

Shortest path analysis finds the path with the minimum cumulative impedance between nodes on a network. The path may connect just two nodes - an origin and a destination - or have specific stops between the nodes.

Closest Facility- finds the closest facility, such as a hospital, fire station, ATM, to any location on a network.

Used by retailers in market studies for siting new facilities

Used by utility companies in managingtheir infrastructure: water, sewer, power

Used by consumers to get directions

Used by agencies to map out service areas: fire, police, public transportation facilities.

Three components:

-geometric model: (x,y,z) coordinates of edges and junctions

-logical model: which edges are connected to what junctions

-addressing model: location on the network using measure

A geometric network is a connectivity relationship between a collection of feature classes in a feature dataset.

Each feature has a role in the geometric network of eitehr an edge or a junction.

Multiple feature classes may have the same role in a single geometric network.

Network feature classes: junctions and edges

Edges can be

-simple: one attribute record for a single edge

complex: one attribute record for several edges in a linear sequence

JUNCTIONS

Junctions exist at all points where edges join

-If necessary they are added during network building (generic junctions)

Junctions can be placed on the interior of an edge (e.g. stream gage)

Any number of point feature clases can be built into junctions on a single network.

Each junction feature class in a network can have junctions which are sources or sinks for flow

Flags:

-define the starting points for traces

-an upstream trace, use a flag to specify where the upstream trace will begin

-flags can be placed aanywhere along edges or on junctions

Barriers:

-define places in the network past which traces cannot continue

-Interested in tracing on a particular part of your network, you can use barriers to isolate that part of the network.

-like flags, barriers can be placed anywhere along edges or junctions

Hydrologic data includes:

-single-line streams

-double-line streams

-braided streams

-manmade channel systems

-waterbodies

Network Building

-define flow-paths within double-line streams and waterbodies

-define network sinks and sources

Network Flow Direction

-enable flow in flow-paths

-disable flow ni shorelines

Network tracing

-can trace upstream or downstream

-can find the shortest path b/w two points on the network

Can connect drainage areas to the Network (area goes to point on line)

Addressing: name the point on the line

-relative (% along length)

-or absolute (distance)

Proportional Aliasing:

-Distance is measured relative to the length of the line as a percentage (0-100)

Allocation measures the efficiency of public facilities, such as fire stations, or school resources, in terms of their service areas

Location-allocation solves problems of matching the supply and demand by using sets of objectives and constraints

Facility role

-what rolee does each facility play?

-What processes are performed at each facility?

Facility locations:

-Where shoudl they be located to meet strategic objectives

Capacity allocation:

-How is capacity allocated to each facility

-How is capacity utilized within each facility?

Market and supply allocation

-What markets should each facility serve?

-Which supply sources should feed each facility?

Types of Facilities:

-Offshore facility: low cost for export production (produce in anotehr country for "export" to specific country)

-Source Facility: low cost for global production (expands offshore to entire global network - requires higher skilled workforce)

-Server facility: regional production facility - serve the market where it is located (economic incentives)

-Contributor Facility: server facility with development skills - serves local market, but can develop products to suit customer needs

-Outpost facility: regional facility targeting acquisition of local skills (e.g. plants located in Germany

-Lead facility: creates new products and/or process technologies

Technological

Macroeconomic

-tariffs and tax incentives

-exchange rate and demand risk

Political

-political stability

-independent legal systems (protects investment)

Infrastructure

-availability of land for sites

-level of skilled labour

-access to transportation (rail service, higheways, airports, seaports), local utilities.

Competitive

Positive externalities: co-location of multiple firms benefits each of the firms - increases overall demand

Contributes to infrastructure development

No positive externalities: seek to split the market (maximize share) - local facilities close together (compete on distance rather than price)

Logistics and faciliti costs: must consider inventory, transportation, and facility costs

-both costs increase as # facilities increase

-transportation costs decrease as # facilities increase

Phase I: define a supply chain strategy

-define a competitive strategy (how the supply chain aims to satisfy customer needs)

-specify teh network that will best support the competitive strategy

-forecast evolution of global competition

-determine if competitors will be local or global

-identify constraints on available capital

-determine if growth will be accomplished by acquisition, building, or partnering.

Phase II: define the regional facility configuration

-Forecase demand by country

---assessment of size of demand

---determination of variance in customer requirements across borders (homogeneous = large faciliy)

-Identify economies of scale or scope

---if significant, few facilities serving many markets

-If not, distribute facilities across markets

-Identify demand, exchange-rate, and political risk with different regional markets

---include tariffs, requirements for local production, tax incentives, etc.

---locate where facility can keep majority of profits with lowest tax rates

-Identify competitors in each region

---determine if facilities should be close to competitors

---identify desired response time for each market and logistics cost

-Configuration defines teh number of facilities, regions to be located, and number of markets served by each facility.

Phase III: Select set of desirable potential sites based on:

-infrastructure availability to support desired production methods

-hard infrastructure: availability of suppliers, transportation services, communication, utilities, warehousing, etc.

-Soft Infrastructure: availability of skilled workforce, workforce turnover, community receptivity.

Phase IV:

-select precise location and capacity allocation for each facility

-design to maximize total profits.

Allocation is the modeling of resourec distribution through a spatial network and teh determination of service zone

In ArcGIS command allocate assigns network links to centres

Based on available supply at centres and the demand assocaited with the links

Links are assigned to centres along the lease impedance path.

When a link is asigned to a centre the available supply at the centre is diminished by teh link demand

The allocation ceases when teh centre supply is exhausted or when teh allocation max distance is exceeded.

Allocating demand to production facilities

Locating facilities and allocating capacity

Key Costs:

-fixed facility cost

-transportation cost

-production cost

-inventory cost

-coordination cost

Which plants to establish? How to configure the network?

IDEAL INFORMATION AVAILABLE

-Location of supply sources and markets

-Location of potential sites

-Demand forecase by market

-Facility, labour, and material costs by site

-Transportation costs between each pari of sites

-sale price of product by region

-taxes and tariffs

-desired response time

-other service factors

The capacitated plant location model (Phase II)

-Focus on minimizing the cost of meeting global demand

-Includes:

---# potential plants

---# of markets/demand

---annual demand from each market

---potential capacity of each plant

---annualized fixed cost of keeping factory open

---cost of producing and shipping one unit from a specified factory to a specified market (production, inventory, transportation, tariffs)

-Model produces configuration that minimizes overall cost

Gravity Location Models (Phase III)

0Used to find locations that minimize the cost of trasnporting both raw materials from suppliers and finished goods to markets served

-Relevant costs:

---cost of shipping one unit one mile from supply source to facility

---cost of shipping one unit one mile from facility to market

-Location coordinates

-quanitity in tons

-model produces coordinates of locations that minimize the total cost of transporting between supply sourec and facility and b/w supply sourec and market

Phase IV: Allocating capacity to plants and plants to markerts

1.GIS is a tool that can process, display, and integrate different data sources including maps, digital elevation models (DEMs), GPS (Global Positioning System), images, and tables.
2.Models built with a GIS can be vector-based or raster-based.
3.The distinction between raster-based and vector-based models does not preclude GIS users from integrating both types of data in the modeling process.
4.The process of modeling may take place in a GIS or require linkingof GIS to other computer programs.

GIS and Other Modeling Programs

-loose coupling

-tight coupling

-embedded system

GIS is used to to retrieve and preprocess the spatial data into the form required by the model structure
•Data are written to files that are then used as input to an external modeling tool, which may reside on another computer
•The tool computes the results and returns them as files of data that are then displayed in the GIS

Strengths
–Little knowledge of programming is necessary
–Often quick, if the model only needs to be run once
–Portable –modeling tool can be used with different GIS
Weaknesses
–May involve considerable work in changing data formats and data structures
–Potentially time consuming and tedious if model needs to be run on many different occasions

Model is developed outside GIS
–Has its own data structure with data exchange between the model and GIS hidden from the user.
Exchange of data between GIS to the model for processing and display of results
–Input routines are written so that they are addressed directly by the interactive tools of GIS
–Exchange of data is automatic
Requires considerable investment in programming and data management

Portability is restricted
Examples include hydrologic models such as SWAT and ANSWERS
–The Soil and Water Assessment Tool
–Areal Non-point Source Watershed Environmental Response Simulation model

Microsoft supports a special clinet/server mechanism called dynamic data exchange (DDE)

DDE enables two applications to interact by exchanging data

ArcGIS supports DDE and can communicate with any other application that supports DDE such as Visual Basic, Excel, Access

DDE does have one limitation: the applications must be running the same computer (DDE does not support networks)

Model is written using the analytical engine of the GIS (API, scripting tools, map algebra)

Or, specific GIS functions such as overlay, spatial data query, and display are added to a complex modeling system

Current GIS tools provide many logical and mathematical functions for model development inside GIS

Advantages
–Don’t have to transform data to other formats
–Data structures do not have to be matched
–Users are able to make on-the-fly changes more rapidly

Considerations
–If good models already exist, just use loose coupling or tight coupling

Binary

Index

Regression

Process

A binary modeluses logical expressions to select map features from a composite map or multiple grids. The output of a binary model is in binary format: 1 (True) for map features that meet the selection criteria and 0 (False) for map features that do not.

Can be vector or raster based

An index model calculates the index value for each unit area and produces a ranked map based on the index values.

An index model is similar to a binary model in that both involve multi-criteria evaluation and both depend on map overlay operations in data processing.

An index model differsfrom a binary model in that it produces for each unit area an index value rather than a simple yes or no.

CRITERION WEIGHTS

Criterion weight refers to the relative importance of a criterion as compared to other criteria.
Many studies have used expert-derived paired comparisons for evaluating criteria.
Paired comparison is available in various software packages (e.g., Expert Choice).

Data may need to be standardized, simplified, or aggregated

Terrain Stability Mapping Using the SINMAP extension.
The SINMAP extension was developed by TerratechConsulting Ltd., Utah State University, and Canadian Forest Products.
Funding was provided by Forest Renewal B.C., Canada.

Based on an objective physical model.
Can be quantitative or qualitative.
Is expressed in terms of stability indices.
Can be combined with geomorphic and statistical analysis.

Where is it used?

Forest and watershed management, forestry and forest engineering.
Determine volume of harvestable timber in annual allowable cut calculations.
Better plan timber development to minimize occurrence of landslides and resulting impacts.

Use a slope map..

Catchment area map...

Get a SINDEX map..

CONCLUSIONS

•Extended the infinite plane slope stability model with topographically based wetness to account for parameter uncertainty.
•Successfully mapped potentially unstable terrain. 56% of landslides in 8.5% of area classified as “upper threshold”.
•Implementation is an interactive ArcGIS extension, named SINMAP.

A regression model relates a dependent variable to an umber of independent (explanatory) variables in an equation, which can then be used for prediciton or estimation. Tehre are two types of regression model: linear regression and logistic regression.

y = a + b1x1 + b2x2 ... + bnxn

...where y is teh dependent variable, x is teh ind, and b are the regression coefficients.

-all variables are numeric

ESTIMATING ANNUAL FLOOD DISCHARGE WITH REGRESSION MODEL

Q=aAbBc...Nn, or

logQ=loga+b(logA)+c(logB)+...n(logN)

Where Q = flood discharge

A,B,..N = basin characteristics

a = regression constant, and b,c,..n = regression coefficients

Basin characteristics include contributing drainage area, channel slope, stream length, basin elevation, storage, forest cover, annual precipitation, precip intensity, and minimum January temp, etc.

LOGISTIC REGRESSION

-teh log of the odds (probability divided by one minus the prob) of the outcome is modelled as a linear function of the explanatory variables

A process model integrates existing knowledge about the environmental processes in teh real world into a set of relationships and equations for quantifying the processes

Modules or sub-modules are often needed to cover different components of a process model

Some of the modules may use mathematical equations derived from empirical data, wile others may use equations derived from laws of physics

RUSLE

Revised Universal Soil Loss Equation

A = RKLSCP

A= average soil looss

R= rainfall runoff erosivity factot

K= soil erodibility factor

L= slope length factor

S= sleope steepness

C= crop management

P = support practice

Can be used to compute the effects of soil loss management practices

Use x to predict Y

The difference b/w a data point and the regression line is called the residual

The best fitting line minimizes the residuals for all the points, and makes their sum zero.

Non-Spatial... R-squared value looks at the deviations from the regression line; data patterns about the regression line.

Need to look at data! Can have clumps or outliers or nonlinear trends

MAP VARIABLes

The dependent map variable is the one that you want to predict from a set of existing or easily measured variables (independent)

-e.g. home cost with various loan concentrations, housing density, home age, etc.

MAP REGRESSION RESULTS

Scatter plots are regression equations relating Loan Density to three candidate driving variables (Housing Density, Value and Age)

The "R-Squared Index" provides a general measure of how good the predictions ought to be - 40%, 46% indicates moderately weak predictors.

GENERATING A MULTIVARIATE REGRESSION

A regression equation using all three independent map variables using multiple linear regression is used to generate a prediction map.

EVALUATING MULTIPLE LINEAR REGRESSION RESULTS

A regression equation using all independent variables. Used to generate a prediction map. Compared to the actual dependent variable data - Error Surface.

USING THE ERROR MAP TO STRATIFY

Stratify the data set by breaking it into groups of similar characteristics, then generate separate regressions.

-The error surface is used to identify locations where the general regression equation is performing poorly

Other stratification techniques include indigenous knowledge, level-slicing, and clustering

Preprocessing involves conversion of raw data into consisten units that accurately represent mapped conditions.

1. Calibration - ensuring that an instrument or measurement conforms to a standard

2. Translation - converts map values into appropriate units for analysis, such as feet into metres or bushels per acre (measure of volume) into tons per hectare (mass)

3. Adjustment/Correction - Dramatically changes the data, such as post processing GPS coordinates and/or Mass Flow Lag adjustment

4. Normalizing Mapped Data

-involves standardization of a data set, usually for comparison among different types of data

-applying normalization generates a standardized map based on an ideal or goal. This map can be used in analysis with other goal-normalized maps, een from different types of data (because it is now normally distributed)

The "scan" operation moves a window around the yield map and calculates teh coefficient of variation with a 2-cell radius of each location

... higher values indicate areas with more localized variability

Stratification partitions the data (numeric) or the project area (spatial) into logical groups-

Proximity map identifies the distance from point, line, or polygon feature to all otehr locations

How similar are two maps?

Thematic Categorization

-We often represent continuous spatial data (map surfaces) as a set of discrete polygons.

-Which classified map is correct?

How similar are the three maps?

Spatial Precision

(where-boundaries)

-of points, lines and areas is a primary concern of GIS, but we are often less concerned with Thematic Accuracy (what - map values)

Use contingency tables to determine coincidence
(make categories the same for both maps, and see how often then are coincident)

Proximal Alignment

-where are the misses?

-proximal alignment isolates a category on one of the maps, generates its proximity, then identifies the proximity values that align with the same category ont he other map

Three ways to compare map surfaces...

-statistical tests

-percent difference

-surface configuration

Statistical tests compare one set of cell values to that of another based on the differences in teh distributions of the data - 1) data sets (partition or coincidenec; continuous or sampled) 2) statistical procedure (t-Test, f-Test, etc)

Percent Difference capitalizes on the spatial arrangement of thevalues by comparing the values at each map location

Surfae Configuration

-capitalizes on teh spatial arrangement of teh values by comparing the localized trend in teh values - Slope Map, Aspect Map, Surface Configuration Index

Temporal Difference

-Map variables... map values within an analysis grid can be mathematically and statistically analyzed (take the difference between two times)

Spatial Variable Dependence - what occurs at a location in geographci space is related to:

-the conditions of that variable at nearby locations, termed Spatial Autocorrelation (intra-variable dependence)

-the conditions of other variables at that locations, termed Spatial Correlation (inter-variable dependence)

Map Stack - relationships among maps are investigated by aligning grid maps with a common configuration... #cols/rows, cell size and geo reference.

Data Shishkebab- each map represents a variable, each grid space a case and each value a measurement with all of the rights

Isolate areas with mean +1 StDev (tail of normal curve)

-make them a category

Sum of a binary progression (1, 2, 4, 8..) provides level slice solutions for many map layers

Can locate combinations of selected multivariate measurements in geographic space (ie. high 1,2,3; high1,2,not 3; etc)

CALCULATING DATA DISTANCE

An n-dimensional plot depicts the multivariate distribution; the distance between points determines the relative similarity in data patterns

Geographic space- relative spatial position of measurements (spatial distribution)

Data space- relative numerical position of measurements (numerical distribution)

The relative data distance between teh comparison point's data pattern and those of all other map locations form a Similarity Index

(question 5?)

A data cluster is a zone in the field with similar levels of a number of variables.

Groups of "floating balls" in data space identify locations in teh field with similar data patterns - data zones.

Fertilization rates vary for the different clusters (if clusters are P, K, and N levels)

ASSESSING CLUSTERING RESULTS

Clustering results can be roughtly evaluated using basic statistics.

-mean, SD, Min, max within each cluster are calculated. Ideally the mean b/w two clusters would be radically different and the SD small to large differences between groups and small differences within groups

-can use Box and Whisker Plots to visualize differences

HOW CLUSTERING WORKS (e.g. height vs weight)

1) you ahve a scatterplot (x/y)

2) the data distance to each weight/height measurement pair is calculated and the point is assigned to the closest arbitrary cluster centre

3)the average x,y coordinates of the assigned students is calculated and used to reposition the cluster centres

4)Repeat data distances, cluster assignments and repositioning until no change in cluster membership (centres do not move)

Spatial Data Mining (the big picture)... making sense out of a map

The Spatial Data Mining Process

-geo-registered maps are used to uncover and apply spatial relationships

Maps of the item of interestand related variables are encoded then analyzed to derived a "map-ematical" relationship that can be used to predict the item at another place or time.

For example, test market sales of a product can be related to demographic characteristics then the derived relationsips moved to another city to generate a map of predicted sales.

DATA MINING

-Mapped data that exhibits high spatial dependency create strong prediciton functions. As in traditional statistical analysis, spatial relationships can be used to predict outcomes.

-The difference is that spatial statistics predict where responses will be high or low.

Spatial data mining operations involve characterizing numerical patterns and relationships among mapped data.

Geotechnology > GIS > Map Analysis > Spatial Statistics > Spatial Data Mining

-Interrelationships within and among map layers

CAN BE:

Descriptive (summer statistics, comparison, classification, e.g. clustering)

Predictive (math/stat relationships among map layers e.g. regression)

Prescriptive (appropriate actions, e.g. optimization)

Four primary types of applied spatial models:

1)Suitability - mapping preferences (e.g. habitat and routing)

2)Economic - mapping financial interactions (e.g. combat zone and sales propensity)

3)Physical - mapping landscape interactions (e.g. terrain analysis and sediment loading)

4)Mathematical/Statistical - mapping numerical relationships

Descriptive math/stat models summarize existing mapped data

-standard normal variable map for unusual conditions and clustering for Data Zones

Predictive math/stat models develop equations relating mapped data

-map regression for equity loan prediction and probability of product sales

Prescriptive math/stat models identify management actions based on descriptive/predictive relationships

-retail marketing and precision Ag

______

Discrete Actions: If <conditions> then <Actions>

-ie. if P is ___ then apply ___ fertilizer

Computer Mapping automates the cartographic process (70s)

Spatial DB managemnet links computer mapping techniques with traditional db capabiltities (80s)

Map Analysis representation of relationships within and among mapped data (90s)

Multimedia mapping full integration of GIS, Internet, and visualization technologies (00s)

Traditional GIS > Spatial analysis

(features, discrete objects, mapping) > (cells, surfaces, continuous geographic space, contextual spatial relationships)

Traditional Statistics > Spatial Statistics

(mean, STDEV, central tendency) > (map of variance, spatial distribution surface, numerical spatial relationships)

GIS

-70s- computer mapping

-80s-spatial db managemnt

-90s Map analysis

---Spatial Analysis (geographic context)

---Spatial statistics (numeric context)

SPATIAL ANALYSIS

-reclassify

-overlay

-proximity

-neighbours

SPATIAL STATISTICS

-surface modeling (point data to continuous spatial distributions)

-Spatial Data Mining (interrelationships within and among map layers)

Image Draping - is an established technique in GIS. Draping a topographic or thematic map onto a 3D terrain surface is effective but relies on abstract colors, shading, and symbols.

Geometric Modeling - builds a scene of realistic features, such as trees, from a given viewpoint based on GIS inventories of tree types and densities

Rendering Technique

1) 3D terrain surface

2) polygon containers

3)surface texture

4) tree objects

5) final compositions

6) atmospheric effects

Can visualize before/after fire or avalanche

Contemporary visualization software is incredible

From Data to Analysis

Past: small amount analysis, huge data conversion

Future: huge spatial analysis, small data conversions (ubiquitous data)

From description to simulation and modelling

PAST

Iconic models: scaled down representations of the real thing.

FUTURE

Symbolic models: based on logical relationships in mathematical or statistical form

From 2D description to 4D interaction

PAST

2D flat map displays

FUTURE

Effective 3D visualization

4D incorporation of time

5/6/7D incorporation of touch (pressure, texture, temperature), sound and smell

User as participant

out of this world

past = GIS applied to earth

future = GIS as a methodology for the analysis of spheres (other planets, the human brain)

GIServices

-location based (for organizatons and individuals)

---facilities, markets, where am I? etc

-Content tailored for current locations ,not the desktop

GISpecialists

geospatial information scientists, specialists (or students)

-appreciative of teh braod ranging, integrative role geospatial data can play

-highly knowledgeable with respect to the unique challenges of geospatial data

-fully conversant with mainstream information technology

-sufficient expertise in an application area (geology, local gov, marketing, etc) to make a contribution

DATA IS STILL AT THE HEART!

Dominant IT issue in 200s is data

-was hardware 70s/80s

-software 80s/90s

Not an issue of acquiring, but of managing

Will availability be..

-plentiful and  cheap?

-in infinite detail if you can afford it?

-severely curtailed by legal controls to ensure personal privacy?

TRAGEDY OF COMMONS IN MAKING?

Multiple individuals, acting independently and rationally consulting their own self-interest, will ultimately deplete a shared limited resource even when it is clear that it is not in anyone's long-term interest for this to happen.
Invasions of privacy through detailed data collection and its pervasive distribution produces a backlash of demand for privacy

The expense of data production, but the ease of re-production and distribution, reduces the value of data to zero and chokes off its availability
–Is public domain data the information age equivalent of the agricultural commons?

Or is the future a self generating system? (more GIS use > better decisions > more investment > better data > more GIS use)

global/local

exact/inexact

deterministic/stochastic (error)

GLOBAL

trend (determ/inexact)

regresstion (stoch/exact)

LOCAL

thiessen (det/ex)

density (det/inex)

IDW (det/exact)

splines (det/ exact)

Kriging (stoch/exact)