Uncertainty Analysis and DataAssimilation of Remote Sensing Data forthe Calibration of Cellular AutomataBased Land-Use Mod...
Introduction» Land-use change models are becoming important instruments for the  assessment of policies aimed at   » impro...
MOLAND land-use model for Dublin                   13/07/2012      3
MOLAND land-use model for Dublin                                   Land use                                               ...
Historic calibration     » Land-use change models are typically calibrated using a historic       calibrationModel initial...
Land-use data for calibration» Dynamic land-use change models require for their calibration time series  of high quality a...
Remote sensing data for calibration                       1994             1997                 New image                 ...
Spatial Metrics» Spatial metrics:   » Quantitative measures to describe structures and patterns in the      landscape» Cal...
Uncertainties in predicted land use» A major shortcoming in the historic calibration of land-use change models  is that un...
Spatial metric                  Inferred land                  use Image interpretation                                   ...
Objectives» Main objectives of the Belspo STEREO II ASIMUD project:   » Improve land-use simulations: lower uncertainties ...
Calibration with data-assimilation algorithm» Data-assimilation algorithms   » integrate observations of the state of a sy...
Workflow1.   Model in error propagation mode     (Monte Carlo simulations)     » Uncertain model parameters2.   Model in d...
Simplified MOLAND land-use model for DublinSimplifiedland-use       Original MOLAND land-usemodel                   catego...
Simplified MOLAND land-use model for Dublin» Neighbourhood influence rules: 5 parameters  2 parameters  (exponential func...
Quantification of uncertain input parameters                                                 Sill (s)                   Ra...
1. Error propagation - Probability maps              1990      1997      2001         2006     2010    Employment related ...
1. Error propagation - Spatial metricsPD – Patch DensityNumber of urban patches (patches/100ha)                           ...
1. Error propagation – Spatial metricsPLADJ – Percentage of LikeAdjacencies                                        PLADJ =...
2. Data assimilation – Particle FilterStep A:» Apply Bayes’ equation to   realizations of the model» Results in a ‘weight’...
Step 1: Apply Bayes’ equation to each realization (particle) i                       Prior: PDF of                        ...
Calculating weights                            æ 1é                                  (i) ùö    (             )   = exp ç- ...
2. Data assimilation - Particle filter               Observations           Observations                         13/07/201...
2. Data assimilation - Particle filter Number of copies or clones199720012006                                           Po...
2. Probability maps with data assimilation              1990      1997      2001        2006     2010    Employment relate...
Conclusions» Monte Carlo framework for error propagation modelling and particle  filtering was applied to a simplified ver...
http://www.asimud.be        13/07/2012     27
Upcoming SlideShare
Loading in …5
×

Uncertainty Analysis and Data Assimilation of Remote Sensing Data for the Calibration of Cellular Automata Based Land-Use Models

944 views
822 views

Published on

Preliminary results of the Belspo STEREO II project presented at the iEMSs 2012 conference from 1 - 5 July in Leipzig

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
944
On SlideShare
0
From Embeds
0
Number of Embeds
32
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • The historic calibration is typically done with land-use maps with a ten years interval as indicated in the figure. The reason is that production of land-use maps is elaborate and time-consuming, because it is usually based on visual interpretation of remote sensing data in combination with other datasets. This also leads to temporal inconsistencies. The sporadic availability and temporal inconsistencies hamper the historic calibration of land use change models
  • Long timeseries of MR remote sensing images. Explain that a method is being developed that uses spatial metrics that describe characteristic aspects of urban form and structure. Parameters in the model are tuned in such a way that the simulated patterns of urban growth, as described by the metrics, match the patterns observed in remote sensing imagery
  • At the timestep of satellite overpass the value of the indicator (spatial metric) and its uncertainty needs to be weighted in order to estimate the optimal model parameters at this timestep
  • Original Moland model: 8 functions * 23 land-use classes  184 possibleinteractionrules * 5 parameters  920 possible parametersSimplifiedversion: 2 functions * 4 land-use classes  8 possibleinteractionrules, only 4 are taken into account in calibrationscheme * 2 parameters  8 possible parameters to calibrate
  • AGGREGATION METRICPLADJ increases lessfragmentation/more aggregationintolargerpatches
  • Announcethat we plan toanalyse uncertaintypropagation in different metricsAssimilationusingvectors of metrics/land-usecombinations as filter variablesApply the framework to the f
  • Uncertainty Analysis and Data Assimilation of Remote Sensing Data for the Calibration of Cellular Automata Based Land-Use Models

    1. 1. Uncertainty Analysis and DataAssimilation of Remote Sensing Data forthe Calibration of Cellular AutomataBased Land-Use ModelsJohannes van der Kwast UNESCO-IHE, the NetherlandsLien Poelmans, Inge Uljee, Guy Engelen VITO, BelgiumTim Van de Voorde, Casper Cockx, Frank Canters Vrije Universiteit Brussel, BelgiumKor de Jong, Derek Karssenberg Utrecht University, the Netherlands
    2. 2. Introduction» Land-use change models are becoming important instruments for the assessment of policies aimed at » improved spatial planning » sustainable development » scenario analysis» Need for robust and more reliable tools» Correct calibration and validation of land-use change models is of major importance 13/07/2012 2
    3. 3. MOLAND land-use model for Dublin 13/07/2012 3
    4. 4. MOLAND land-use model for Dublin Land use Land use & Interaction Stochastic at time T+1 weights perturbation t v 1 ln rand Suitability & 0 0.5 1& Transition Rule Change cells to land use for Time Loop which they have the highest transition potential until the demands are met. Transition Accessibility Zoning Potentials & & = 13/07/2012 4
    5. 5. Historic calibration » Land-use change models are typically calibrated using a historic calibrationModel initialisation Hindcast Forecast 1990 2000 2030 not Ok OkActual map 1990 Actual map 2000 parameters Courtesy of EC JRC 13/07/2012 5
    6. 6. Land-use data for calibration» Dynamic land-use change models require for their calibration time series of high quality and consistent land-use information.» Remote sensing data can be used to » Correct inconsistencies in land-use maps available for calibration » Produce land-use information at more time steps » Provide additional land-use information to improve calibration 13/07/2012 6
    7. 7. Remote sensing data for calibration 1994 1997 New image RS New RS Data RS Data DataModel initialisation Hindcast Forecast 1990 1994 1997 2000 2030 Modelsimulation Source: MAMUD projectActual map 1990 Actual 13/07/2012 map 2000 7
    8. 8. Spatial Metrics» Spatial metrics: » Quantitative measures to describe structures and patterns in the landscape» Calculation at different levels of abstraction, e.g. patch, class, moving window or landscape scale» Examples of spatial metrics are: fractal dimension, contagion, edge density, patch density, adjacency event 13/07/2012 8
    9. 9. Uncertainties in predicted land use» A major shortcoming in the historic calibration of land-use change models is that uncertainties are neglected. Uncertainties mostly exist in: » Model parameters » Reference data used for calibration of the model» This leads to uncertainties in the prediction of land use 1994 1997 New image RS New RS Data RS Data Data 1994 1997 13/07/2012 9
    10. 10. Spatial metric Inferred land use Image interpretation Calibrated model parametersModelinitiation Predicted land use Spatial metric 13/07/2012 10
    11. 11. Objectives» Main objectives of the Belspo STEREO II ASIMUD project: » Improve land-use simulations: lower uncertainties compared to other automatic calibration methods » Development of an automatic calibration method using remote sensing data in an innovative data-assimilation approach » Robust and reliable tools for land-use change modelling and calibration for use in policy contexts will be facilitated and promoted » The probability maps of simulated land use will be valuable additional data for end users to assess planning policies 13/07/2012 11
    12. 12. Calibration with data-assimilation algorithm» Data-assimilation algorithms » integrate observations of the state of a system with the modelled state (the hindcast) to produce the best estimate of the parameter values and state variables. » balance the uncertainty in the observation data and in the hindcast. » provide calibrated parameters as probability distributions» We apply the Particle Filter, a robust Monte Carlo based method, implemented in a Python framework» Data assimilation is often used in atmospheric chemistry models, weather forecasting, hydrological modelling, GPS technology and astronomy» Relatively new in the field of land-use change modelling 13/07/2012 12
    13. 13. Workflow1. Model in error propagation mode (Monte Carlo simulations) » Uncertain model parameters2. Model in data assimilation mode (Particle Filter) » Uncertain observations Observations Observations 13/07/2012 13
    14. 14. Simplified MOLAND land-use model for DublinSimplifiedland-use Original MOLAND land-usemodel categories Residential continuous dense urban fabric, ResidentialPopulation continuous medium denserelated urban fabric, Residentialclasses discontinuous urban fabric, Residential discontinuous sparse urban fabricEmployment Industrial areas, Commercialrelated areas, Public and privateclasses services, Port areas, Arable land, Pastures, Forests, Semi-natural areas, Wetlands,Non urban Artificial non-agricultural vegetated areas, Construction sites Road and rail networks and associated land, Abandonment,Other Mineral extraction sites, Airport, Water bodies, Restricted access areas, Dump sites 13/07/2012 14
    15. 15. Simplified MOLAND land-use model for Dublin» Neighbourhood influence rules: 5 parameters  2 parameters (exponential function) (1, a) (b, c) (d, 0) (0, inertia) 13/07/2012 15
    16. 16. Quantification of uncertain input parameters Sill (s) Range (r)From To Mean Mean min (SD) max min (SD) max 50.5 0.41Population Population 1 (25) 100 0.12 (0.2) 0.7 -25.0 0.205Population Employment -100 (25) 50 0.01 (0.2) 0.4 -50.0 0.355Employment Population -100 (25) 0 0.01 (0.2) 0.7 50.5 0.455Employment Employment 1 (25) 100 0.16 (0.2) 0.75 Range = 0.41 Sill Sill Range = 0.7 Range = 0.12 13/07/2012 16
    17. 17. 1. Error propagation - Probability maps 1990 1997 2001 2006 2010 Employment related urban Population related urban 13/07/2012 17
    18. 18. 1. Error propagation - Spatial metricsPD – Patch DensityNumber of urban patches (patches/100ha) 13/07/2012 18
    19. 19. 1. Error propagation – Spatial metricsPLADJ – Percentage of LikeAdjacencies PLADJ = 0 PLADJ = 65Degree of aggregation of the urban patchesPLADJ = 0: urbanised area is maximally disaggregatedPLADJ = 100: one large urban patch Population class Employment class 13/07/2012 19
    20. 20. 2. Data assimilation – Particle FilterStep A:» Apply Bayes’ equation to realizations of the model» Results in a ‘weight’ assigned to each realizationStep B:» Clone each realization a Step B number of times proportional to Step A the weight of the realization 13/07/2012 20
    21. 21. Step 1: Apply Bayes’ equation to each realization (particle) i Prior: PDF of Prior: PDF of model observations realization i Prior: PDF of observations given the model realization i Posterior: probability distribution function (PDF) of realization i given the observations 13/07/2012 21
    22. 22. Calculating weights æ 1é (i) ùö ( ) = exp ç- ëyt - Ht ( x t )û Rt ëy t - Ht ( x t )û÷ (i) ù -1 é T p yt x (i) t è 2 ø Weight of particle Measurement operator = 1 Measurement error Model variance Median value realization and covariance of of spatial observations metrics for observations at time step t 13/07/2012 22
    23. 23. 2. Data assimilation - Particle filter Observations Observations 13/07/2012 23
    24. 24. 2. Data assimilation - Particle filter Number of copies or clones199720012006 Population Class 13/07/2012 24
    25. 25. 2. Probability maps with data assimilation 1990 1997 2001 2006 2010 Employment related urban Population related urban 13/07/2012 25
    26. 26. Conclusions» Monte Carlo framework for error propagation modelling and particle filtering was applied to a simplified version of the MOLAND model for Dublin» First results seem promising» However, relative big gap between spatial metrics calculated from the RS- based land-use maps and the MOLAND land-use map may hamper the analysis 13/07/2012 26
    27. 27. http://www.asimud.be 13/07/2012 27

    ×