This time-series land-use/land-cover (LULC) classification data set is derived from Landsat multispectral data for the Upper Peace Region of Alberta, captured between 1985 and 2015.
This land classification raster dataset contains 11 classes: 0 - Unclassified, 1 - Shrubland, 2 - Grassland, 3 - Coniferous Forest, 4 - Deciduous Forest, 5 - Open Water, 6 - Wetland (Bog/Fen/Marsh/Swamp), 7 - Exposed Land/Bare Soil, 8 - Cropland/Pasture, 9 - Developed Footprints, and 10 - Fire Scar.
The time-series are stacked to create a 7-band TIFF image with each band representing a five-year interval between 1985 and 2015.
The Overall accuracy is the total number of correctly classified pixels divided by the total number of pixels in the error matrix N.
The Producer accuracy is an indication of the probability of a reference pixel being correctly classified. It is calculated by dividing the number of correctly classified pixels in a class by the total number of pixels in that class.
The User accuracy is a ratio of the total number of correctly classified pixels in a class divided by the total number of pixels that were classified in that class. It indicates the probability that a pixel classified on the map represents that class on the ground.
The Kappa coefficient indicates if the accuracy level is significantly better than a random result, providing a better comparison of different classifications.
LULC Classification result of 1985
Overall Accuracy = (27906303/34077250) 81.8913%
Kappa Coefficient = 0.7250
Class | Producer Accuracy (%) | User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 30.49 | 20.60 | 109966/360610 | 109966/533906
Grassland | 30.05 | 4.98 | 19462/64768 | 19462/390667
Coniferous | 81.16 | 89.93 | 10953125/13495741 | 10953125/12179655
Deciduous | 86.14 | 90.29 | 12140116/14093066 12140116/13445626
Open Water | 96.53 | 75.29 | 86138/89234 | 86138/114408
Wetland | 76.63 | 61.91 | 4487852/5856823 | 4487852/7249367
Exposed Land | 99.85 | 64.04 | 89196/89327 | 89196/139289
Developed | 73.87 | 84.04 | 20448/27681 | 20448/24332
LULC Classification result of 1990
Overall Accuracy = (25606752/30875972) 82.9342%
Kappa Coefficient = 0.7380
Class | Producer Accuracy (%) | User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 37.10 | 20.29 | 108651/292823 | 108651/535510
Grassland | 30.55 | 5.89 | 15356/50257 | 15356/260579
Coniferous | 81.30 | 90.62 | 9961624/12253231 | 9961624/10993227
Deciduous | 87.87 | 90.61 | 11479253/13064264 | 11479253/12668628
Open Water | 89.08 | 41.11 | 11150/12517 | 11150/27121
Wetland | 77.16 | 62.98 | 3944801/5112314 | 3944801/6263318
Exposed Land | 99.59 | 63.35 | 67023/67296 | 67023/105790
Developed | 81.19 | 86.67 | 18894/23270 | 18894/21799
LULC Classification result of 1995
Overall Accuracy = (28333250/34937626) 81.0967%
Kappa Coefficient = 0.7158
Class | Producer Accuracy (% )| User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 46.24 | 36.11 | 302242/653634 | 302242/836950
Grassland | 33.15 | 10.92 | 33523/101133 | 33523/307070
Coniferous | 79.23 | 90.28 | 11009728/13895048 | 11009728/12195353
Deciduous | 86.49 | 90.16 | 12366951/14298720 | 12366951/13716349
Open Water | 96.01 | 74.98 | 99111/103233 | 99111/132190
Wetland | 76.13 | 58.01 | 4336535/5696049 | 4336535/7474929
Exposed Land | 98.16 | 66.56 | 162051/165092 | 162051/243476
Developed | 93.49 | 73.81 | 23109/24717 | 23109/31309
LULC Classification result of 2000
Overall Accuracy = (32407665/40106030) 80.8050%
Kappa Coefficient = 0.7208
Class | Producer Accuracy (%) | User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 54.43 | 52.99 | 756244/1389396 | 756244/1427223
Grassland | 34.45 | 16.65 | 81148/235561 | 81148/487335
Coniferous | 78.91 | 89.54 | 11695821/14822219 | 11695821/13062117
Deciduous | 87.22 | 88.91 | 13943163/15986594 | 13943163/15682247
Open Water | 94.50 | 93.00 | 458742/485452 | 458742/493256
Wetland | 74.52 | 59.80 | 4975817/6676778 | 4975817/8321455
Exposed Land | 99.55 | 77.92 | 442503/444481 | 442503/567877
Developed | 82.73 | 84.05 | 54227/65549 | 54227/64520
LULC Classification result of 2005
Overall Accuracy = (28578934/35752013) 79.9366%
Kappa Coefficient = 0.6956
Class | Producer Accuracy (%) | User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 48.50 | 54.84 | 672097/1385794 | 672097/1225494
Grassland | 30.00 | 11.80 | 57912/193032 | 57912/490933
Coniferous | 77.61 | 91.53 | 11054114/14243226 | 11054114/12076772
Deciduous | 86.80 | 90.19 | 13660454/15738244 | 13660454/15146141
Open Water | 95.19 | 85.34 | 215253/226124 | 215253/252220
Wetland | 72.39 | 43.39 | 2714134/3749484 | 2714134/6255131
Exposed Land | 95.62 | 69.53 | 177467/185597 | 177467/255242
Developed | 90.14 | 54.92 | 27503/30512 | 27503/50080
LULC Classification result of 2010
Overall Accuracy = (29554171/37298849) 79.2361%
Kappa Coefficient = 0.6946
Class | Producer Accuracy (%) | User Accuracy (%) | Producer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 46.79 | 45.15 | 538835/1151546 | 538835/1193517
Grassland | 25.74 | 5.43 | 32737/127188 | 32737/603018
Coniferous | 76.87 | 89.16 | 10412354/13545148 | 10412354/11677713
Deciduous | 85.28 | 89.15 | 13278851/15570587 | 13278851/14895494
Open Water | 96.49 | 76.21 | 202465/209822 | 202465/265682
Wetland | 75.33 | 58.27 | 4856615/6446747 | 4856615/8334262
Exposed Land | 95.93 | 70.02 | 196360/204687 | 196360/280429
Developed | 83.37 | 73.78 | 35954/43124 | 35954/48734
LULC Classification result of 2015
Overall Accuracy = (25010859/30972007) 80.7531%
Kappa Coefficient = 0.7136
Classs | Producer Accuracy (%) |User Accuracy (%) | P roducer Accuracy (Pixel Ratio) | User Accuracy (Pixel Ratio)
Shrubland | 37.84 | 54.20 | 327393/865234 | 327393/604010
Grassland | 21.05 | 5.10 | 17779/84452 | 17779/348315
Coniferous | 78.60 | 89.30 | 9356764/11903527 | 9356764/10477964
Deciduous | 88.00 | 88.49 |10898515/12384004 | 10898515/12315500
Open Water | 95.72 | 82.45 | 78731/82248 | 78731/95485
Wetland | 75.78 | 60.39 | 4103299/5414723 | 4103299/6795049
Exposed Land | 98.35 | 66.06 | 196001/199299 | 196001/296685
Developed | 84.05 | 83.02 | 32377/38520 | 32377/38999
The accuracy assessment of the LULC classification result was conducted using the Digital Integrated Dispositions (DIDs) and Alberta Merged Wetland Inventory (AMWI) data from the Government of Alberta, and the Alberta Biodiversity Monitoring Institute (ABMI) Wall-to-wall Land Cover datasets for 2010.
The Cropland/Pasture class was excluded from accuracy assessment since this class was attributed based on agricultural land data sourced from Agriculture and Agri-Food Canada. In addition, the Fire Scar class was excluded from the accuracy assessment since this class was attributed by matching historical wildfire data sourced from the Government of Alberta that fall within the Exposed Land/Bare Soil class.
Process steps were performed in ENVI 5.4 and GRASS GIS 7.2.2 software to produce the classification map:
1. Pre-release versions of annual Landsat Best Available Pixel Composite (LBAPC) data from 1984 to 2012 with 30 m spatial resolution were obtained from the Natural Resources Canada/Pacific Forestry Centre for testing purposes. Of these, the 1985, 1990, 1995, 2000, 2005 to 2010 data were used as inputs to produce this time-series LULC classification stack. In addition, 2015 LBAPC data were produced in-house using the Google Earth Engine.
2. Ground-reference datasets were created from DIDs and AMWI data and ABMI Wall-to-wall Land Cover data.
3. A Random Forest classification algorithm was applied to the LBAPC data to produce LULC classifications for the Shrubland, Grassland, Coniferous Forest, Deciduous Forest, Open Water, Wetland (Bog/Fen/Marsh/Swamp), and Exposed Land/Bare Soil classes.
4. The Developed Footprint class was produced using the Constraint Energy Minimization and Spectral Angle Mapper partial unmixing method followed by a K-Means clustering. Post-classification techniques (i.e., majority filter, high-pass filter, clump, and sieve) were applied to all of the classes to refine the result and to minimize false detections. False detections (e.g., sandbars and fire scars misclassified as developed footprint) were further reduced by matching the DIDs areas that fall within this class.
5. The Cropland/Pasture class was attributed based on Normalized Difference Vegetation Index values >0.7, contained in agricultural land data obtained from the Agriculture and Agri-Food Canada. In addition, Fire Scar class was attributed by matching historical wildfire data sourced from the Government of Alberta with the Exposed Land/Bare Soil class.
6. Classification maps were refined by keeping the classified areas constant over time for the areas that did not change over time based on the image-to-image changes extracted using the Normalized Difference Built-up Index from 1985 to 2015 with 5-year gap.
Part of the process step was adapted from our previous work published in the following journal paper.
These process steps were adapted from previous work published in the following journal paper.
Chowdhury, S., Chao, D.K., Shipman, T.C. and Wulder, M.A. (2017): Utilization of Landsat data to quantify land-use and land-cover changes related to oil and gas activities in west-central Alberta from 2005 to 2013: GIScience and Remote Sensing, p. 700-720, at http://dx.doi.org/10.1080/15481603.2017.1317453.