These scripts and data are public for a manuscript currently under review.
All datafiles mentioned below can be found on Zenodo: https://doi.org/10.5281/zenodo.5703146
- Download RGB data for lakes and nearby channels: [https://code.earthengine.google.com/66cc087785402ba82d27d49c648a3808]
- Download lake ice data [https://code.earthengine.google.com/11757ed73298fa104aed9c5d5b86f5b6] - by Xiao Yang
- Download of April 2015 and April 2017 lake elevations from ArcticDEM [https://code.earthengine.google.com/73182eaa7b7bf0aa15959dba44bde194]
- Delta lake ice.Rmd - By Xiao Yang, some edits by Wayana Dolan
- description: Using downloaded ice fraction data to model 20-year lake ice phenology
- inputs:
- lakeCoverFraction_1b786b4795b34fa035d55c102cc305e7.csv: Output from step 2 of data download.
- outputs:
- lake_ice_modeled_duration_colville_delta_20210304.RData: Modeled lake ice phenology from Landsat
- NumberOfObsTrans_20210304.RData: number of ice observations from Landsat for each lake within both the breakup and freeze-up periods
Main data processing -- all input files are included in the data repository as are all output files, excluding figure pdfs.
-
1_ColvilleDeltaClassification_updateJuly2021.Rmd
- Description: Intakes lake and channel reflectance data and classifies lakes as high functional connectivity vs low functional connectivity within multiple time periods and using different methods of classificatioj
- input files
- ColvilleDataExport_20211022.csv: Lake reflectance data - from google earth engine script no. 1
- ColvilleChannelExport_1km_20211022.csv, ColvilleChannelExport_2km_20211022.csv", ColvilleChannelExport_5km_20211022.csv", ColvilleChannelExport_10km_20211022.csv: Channel reflectance data from within 1-10km of each lake. From google earth engine script no. 1
- ColvilleShapefilesEdited.shp: The lake polygon shapefiles
- colvilleValidation20200508.csv: Connectivity validation data from GECI imagery
- output files
- colville_dt_20211022: lake connectivity classification results using a decision tree-based approach (the final approach used in the paper analysis)
- colville_rf_20211022: lake connectivity classification results using a random forest-based approach
- colville_pct_20211022: lake connectivity classification results using a percentage-based thresholding technique
- colville_km_20211022: lake connectivity classification results using a k-means-based classification approach
- densityPlotExample.pdf: A pdf of the density plot portion of Figure 2.
- DecisionTreeFigure.pdf: A pdf of the final decision tree (Figure 3).
-
2_ColvilleResultes_updateJuly2021.Rmd
- Description: Uses connectivity classifications to: validate classification methods, study how connectivity has changed over time in the Colville Delta, and analyze the relationship between connectivity and elevation, discharge, and lake ice
- input files
- colville_dt_20211022: lake connectivity classification results using a decision tree-based approach (the final approach used in the analysis)
- colville_rf_20211022: lake connectivity classification results using a random forest-based approach
- colville_pct_20211022: lake connectivity classification results using a percentage-based thresholding technique
- colville_km_20211022: lake connectivity classification results using a k-means-based classification approach
- ColvilleShapefilesEdited.shp: The lake polygon shapefiles
- colvilleValidation20200508.csv: Connectivity validation data from GECI imagery
- colville1992Classification.csv: Connectivity validation data from Jorgenson et al. (1997)
- PilouriasColvilleClassCompareNegBuf_20200520_good.shp: Connectivity validation data from Piliouras and Rowland (2020)
- AprilLakeElev20152017.csv: Mean April lake elevation in 2015 and 2017 from ArcticDEM strip data
- lake_ice_modeled_duration_colville_delta_20210304.RData: Modeled lake ice phenology from Landsat (summary statistics for ice phenology dates)
- lake_ice_modeled_20210304.RData: Modeled lake ice phenology from Landsat (daily modeled ice fraction data)
- NumberOfObsTrans_20210304.RData: Number of ice observations from Landsat for each lake within both the breakup and freeze-up periods
- 150_Alaska_nolakes_widths.shp: Alaskan rivers from the GRWL Summary Statistics database (Allen & Pavelsky, 2018) that are wider than 150m.
- output files
- alaskaMap.pdf: Figure 1 part 1 - map of Alaska
- deltaMap.pdf: Figure 1 part 2 - map of the Colville Delta
- validationFig.pdf: Figure 5 - comparison of connectivity algorithm against three datasets
- timeResultsFigure_map.pdf: The map portion of Figure S1 (connectivity through time)
- timeResultsFigure_plot.pdf: The barplot portion of Figure S1 (connectivity through time)
- temporalSummary_plot.pdf: Figure 6 - connectivity over time
- dischargeBarPlots.pdf: Figure 7 - comparing connectivity within different discharge groups
- elevFig.pdf: Figure 8 - boxplot comparing connectivity and lake surface elevation
- iceFigBarPlot.pdf: Figure 9 - comparison of functional connectivity to lake ice phenology
- iceNoObsPlot.pdf: Figure 4 - number of ice observations used in ice phenology calculations
- iceDayLength.pdf: Figure S2 - a figure comparing lake ice fraction to daylength at Nuiqsut, Alaska
- ColvilleDataExport_20211022.csv: Lake reflectance data - from google earth engine script no. 1
- ID: Lake ID
- system.time_start_mean: The time of the observation--values need to be divided by 1000 before hey can be converted to formal dates
- delta: name of the delta--should always be "colville"
- Blue_mean, Blue_p10, Blue_p90: Blue reflectance for a lake on a given date (either mean, 10%, 90%)--values need to be divided by 1000 to be scaled
- Green_mean, Grean_p10, Green_p90: same as above, for Green reflectance--values need to be divided by 1000 to be scaled
- Red_mean, Red_p10, Red_p90: same as above, for Red reflectance--values need to be divided by 1000 to be scaled
- Gb_ratio_mean, Gb_ratio_p10, Gb_ratio_p90, same as above for Green/Blue reflectance
- Ndssi_mean, Ndssi_p10, Ndssi_p90: same as above, for NDSSI (blue-nir)/(blue+nir)
- Nsmi_mean, Nsmi_p10, Nsmi_p90: same as above, for NSMI (red+green-blue)/(red+green+blue)
- area_m: lake area in m^2
- count: Total number of pekel 90% occurence water pixels within the lake boundary
- Red_count: Total number of cloud free pixels included in the mean,p90, p10 values for all color reflectances
- ColvilleChannelExport_1km_20211022.csv, ColvilleChannelExport_2km_20211022.csv, ColvilleChannelExport_5km_20211022.csv, ColvilleChannelExport_10km_20211022.csv: Channel reflectance data from google earth engine script no.1. Includes average reflectance data for each lake on each date within a 1km, 2km, 5km, or 10km buffer
- Same as above file, except no ID or delta column, and has one additional column, described below
- buffer: distance around the lake to search for channel pixels in kilometers
- ColvilleShapefilesEdited.shp: The lake polygon shapefiles
- ID: Lake ID
- delta: name of the delta--should always be 'colville'
- type: 'delta'--should always be 'delta'
- geometry: the geometry of the lake polygon
- colvilleValidation20200508.csv: Connectivity validation data from GECI imagery
- Row.Number: Index
- ID: Lake ID
- Connected: "y" channel present, "m" uncertain channel presence, "n" no channel presence
- note: any notes made during analysis
- colville_dt_20211022: The results of the functional connectivity classification for each lake within each time period using a decision tree-based method. This is the final classification method used in the final analysis for the paper. RData file. Can be read using read_rds() function in R.
- ID: Lake ID
- time_period: The time period (string) for the classification including temporal periods (“2000-2004”, “2005-2009”, “2010-2014”, “2015-2019”), discharge periods (“1”, “2”, “3”, “4”), and the GECI validation period 2013-2016 (“validation”)
- .pred_class: The predicted functional connectivity of the lake within the specified time period (“connected” or “not connected”)
- split: whether or not the lake (during the validation period) was in the training or testing group. If the observation is from outside the validation period, split = "Neither"
- colville_pct_20211022: The results of the functional connectivity classification for each lake within each time period using a percentage-based method. RData file. Can be read using read_rds() function in R.
- ID: Lake ID
- time_period: The time period (string) for the classification including temporal periods (“2000-2004”, “2005-2009”, “2010-2014”, “2015-2019”), discharge periods (“1”, “2”, “3”, “4”), and the GECI validation period 2013-2016 (“validation”)
- .pred_class: The predicted functional connectivity of the lake within the specified time period (“connected” or “not connected”)
- split: whether or not the lake (during the validation period) was in the training or testing group. If the observation is from outside the validation period, split = "Neither"
- colville_rf_20211022: The results of the functional connectivity classification for each lake within each time period using a random forest-based method.RData file. Can be read using read_rds() function in R.
- ID: Lake ID
- time_period: The time period (string) for the classification including temporal periods (“2000-2004”, “2005-2009”, “2010-2014”, “2015-2019”), discharge periods (“1”, “2”, “3”, “4”), and the GECI validation period 2013-2016 (“validation”)
- .pred_class: The predicted functional connectivity of the lake within the specified time period (“connected” or “not connected”)
- split: whether or not the lake (during the validation period) was in the training or testing group. If the observation is from outside the validation period, split = "Neither"
- colville_km_20211022: The results of the functional connectivity classification for each lake within each time period using a kmeans-based method. RData file. Can be read using read_rds() function in R.
- ID: Lake ID
- time_period: The time period (string) for the classification including temporal periods (“2000-2004”, “2005-2009”, “2010-2014”, “2015-2019”), discharge periods (“1”, “2”, “3”, “4”), and the GECI validation period 2013-2016 (“validation”)
- .pred_class: The predicted functional connectivity of the lake within the specified time period (“connected” or “not connected”)
- type: describes which lake-to-channel band ratio was used for the classification
- colville1992Classification.csv: Connectivity validation data from Jorgenson et al. (1997)
- Row.Number: Index
- ID: Lake ID
- Connected: Connectivity classification from GECI
- MapClassification: Connectivity Classification from Jorgenson et al. (1997)
- PilouriasColvilleClassCompareNegBuf_20200520_good.shp: Connectivity validation data from Piliouras and Rowland (2020)
- delta: Name of the delta--should always be "colville"
- count: number of 'connected' pixels within each lake
- ID: Lake ID
- geometry: lake polygons with a negative 30m buffer.
- AprilLakeElev20152017.csv: Mean April lake elevation in 2015 and 2017 from ArcticDEM strip data
- ID: Lake ID
- area_m: lake area in m^2
- count: Total number of pekel 90% occurence water pixels within the lake boundary
- delta: Name of the delta-- should always be "colville"
- elev2015_mean: mean April 2015 lake elevation (m)
- elev2015_median: median April 2015 lake elevation (m)
- elev2017_mean: mean April 2017 lake elevation (m)
- elev2017_median: median April 2017 lake elevation (m)
- lake_ice_modeled_duration_colville_delta_20210304.RData: Modeled lake ice phenology from Landsat
- ID: Lake ID
- ice_duration: modeled total ice duration (days)
- ice_free_duration: modeled ice free duration (days)
- total_transition_duration: modeled total ice transition (breakup and freeze-up periods) duration (days)
- bu_transition_duration: modeled breakup transition duration from breakup start to breakup end (days)
- fu_transition_duration: modeled freeze-up transition duration from freeze-up start to freeze-up end (days)
- buStart: Breakup start--modeled first day of year with ice fraction <80% (day of year)
- buEnd: Breakup end--modeled first day of year with ice fraction <20% (day of year)
- fuStart: Freeze-up start--modeled first day of fall with ice fraction >20% (day of year)
- fuEnd: Freeze-up end--modeled first day of fall/winter with ice fraction >80% (day of year)
- lake_ice_modeled_20210304.RData: Modeled daily ice fraction data from Landsat
- ID: Lake ID
- period: either "BU" for breakup period or "FU" for freeze-up period
- fitted: Modeled ice fraction, ranges from 0 to 1.
- doy: day of the calendar year
- ice_flag: 2 corresponds to ice fractions greater than 0.80, 1 corresponds to ice fractions between 0.20- 0.80, 0 corresponds to ice fractions less than 0.20
- NumberOfObsTrans_20210304.RData: number of ice observations from Landsat for each lake within both the breakup and freeze-up periods
- ID: Lake ID
- count_bu: number of ice observations used one week prior through one week after the modeled breakup period
- count_fu: number of ice observations used one week prior through one week after the modeled freeze-up period
- 150_Alaska_nolakes_widths.shp: Alaskan rivers from the GRWL Summary Statistics database (Allen & Pavelsky, 2018) that are wider than 150m.
- Join_Count: number of point grwl observations contained in the summary grwl reach.
- id: Reach ID (not the same as GRWL reach IDs)
- nchan: number of channels in the reach
- lakeFlag: whether or not the channel is flagged as a lake (should always be 0, which corresponds to not a lake)
- elevm: lake elevation (m)
- widthm: mean reach width (m)
- width_sd: standard deviation in reach width (m)
- width_num: same as widthm
- geometry: reach polyline geometry
- lakeCoverFraction_1b786b4795b34fa035d55c102cc305e7.csv: Output from step 2 of data download of raw lake ice fraction data, included in data repository
- ID: Lake ID
- snowIce: Like ice fraction calculated using Landsat Fmask (fraction, 0-1)
- SLIDE_snowIce: Lake ice fraction calculated using the new SLIDE method from Yang et al. (2021) (fraction, 0-1)
- cloud: Lake cloud fraction (fraction 0-1)
- water: Lake water fraction (fraction, 0-1)
- clear: Lake clear sky/land fraction (fraction, 0-1)
- delta: Name of delta--should always be "colville"
- CLOUD_COVER: Landsat scene total cloudcover (percent, 0-100)
- system:time_start: time of observation (needs to be divided by 1000 to be converted to a date)
- doy: day of year
- LANDSAT_SCENE_ID: Landsat scene ID