Updates to re-pull code

.gitignore

@@ -16,6 +16,9 @@ reports/images/*
 #all tmp folders
 */tmp/*
 
+# pptx files (too big!)
+*.pptx
+
 # Local only folders
 data_local/*
 

README.md

@@ -1,9 +1,11 @@
 # Superior-Plume-Bloom
 
-Description: code to rasterize presence and absence of sediment plumes and algal blooms over time for the western area of Lake Superior
+Description: code to rasterize presence and absence of sediment plumes and algal 
+blooms over time for the western area of Lake Superior
 Contact: B Steele (B dot Steele at colostate dot edu)
 
-This repository is covered by the MIT use license. We request that all downstream uses of this work be available to the public when possible.
+This repository is covered by the MIT use license. We request that all downstream
+uses of this work be available to the public when possible.
 
 Note, there are two methods of authentication used for the Earth Engine workflows.
 A major authentication changed in December 2023 v0.1.383, where authentication 
@@ -12,6 +14,9 @@ was completed using `earthengine authenticate` at the command line. If you are
 attempting to reproduce the code here, but have a version >= v0.1.383, you will
 need to use `ee.Authenticate()` instead of the command line `earthengine authenticate`.
 
+The Methods_Results_Summary.Rmd and associated .html summarize the methodology
+and results of the modeling effort. 
+
 ## Folder Descriptions
 
 *availability_checks*: scripts to document image availability for our AOI
@@ -22,8 +27,10 @@ need to use `ee.Authenticate()` instead of the command line `earthengine authent
 
 *eePlumB*: scripts to develop the 'Earth Engine Plume and Bloom' (eePlumB) labeling modules
 
-*modeling*: scripts to apply supervised kmeans and create output from models
+*modeling*: scripts to work through label data, apply GTB models in GEE, and 
+summarize output
 
 *reports*: Rmd files for report generation
 
 
+

Superior-Plume-Bloom.Rproj

@@ -11,3 +11,6 @@ Encoding: UTF-8
 
 RnwWeave: Sweave
 LaTeX: pdfLaTeX
+
+MarkdownWrap: Column
+MarkdownWrapAtColumn: 80

eePlumB/2_data_ingestion/2_data_re_pull_LS5.Rmd

@@ -7,7 +7,9 @@ output: html_document
 
 # Purpose
 
-This script uses the volunteer point locations exported from `1_data_download.Rmd` which have been manually uploaded as a feature collection to the `ee-ross-superior` project in Earth Engine.
+This script uses the volunteer point locations exported from
+`1_data_download.Rmd` which have been manually uploaded as a feature collection
+to the `ee-ross-superior` project in Earth Engine.
 
 ## R/Python Setup
 
@@ -28,7 +30,8 @@ package_loader <- function(x) {
 invisible(lapply(libs, package_loader))
 ```
 
-Use the conda environment if the env folder is present, otherwise create the environment.
+Use the conda environment if the env folder is present, otherwise create the
+environment.
 
 ```{r conda-env}
 if (!dir.exists("env")) {
@@ -60,7 +63,8 @@ labels = ee.FeatureCollection("projects/ee-ross-superior/assets/labels/collated_
 labels_5 = labels.filter(ee.Filter.eq("mission", "LS5"))
 ```
 
-Get the unique dates of images from this list and assign the date to the `system:time_start` parameter.
+Get the unique dates of images from this list and assign the date to the
+`system:time_start` parameter.
 
 ```{python}
 dates_5 = labels_5.aggregate_array("date").distinct()
@@ -90,7 +94,7 @@ export_l5_meta = (ee.batch.Export.table.toDrive(
   fileNamePrefix = 'LS5_image_metadata',
   fileFormat = 'csv'))
   
-export_l5_meta.start()
+# export_l5_meta.start()
 
 ```
 
@@ -100,6 +104,7 @@ And then apply the scaling factors to the stack
 l5 = (l5
   .map(rp.applyScaleFactors)
   .map(rp.apply_radsat_mask)
+  .map(rp.flag_qa_conf)
   .map(rp.addImageDate))
 
 ```
@@ -113,24 +118,47 @@ for i in range(dates_5.length().getInfo()):
   print(one_date.getInfo())
   one_dt = ee.Date(one_date)
   dt_label = labels_5_dt.filterDate(one_dt, one_dt.advance(1, 'day'))
-  one_image = l5.filterDate(one_dt, one_dt.advance(1, 'day')).mean()
-  
+  one_image_SR_band = (l5
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['SR_B1', 'SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B7'])
+    .mean())
+  one_image_bits = (l5
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 'SR_ATMOS_OPACITY'])
+    .max())
+  one_image_cdist = (l5
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select('ST_CDIST')
+    .min())
+
   #define bands to extract and reduce regions
-  data = (one_image
+  bandsOut = one_image_SR_band.addBands(one_image_bits).addBands(one_image_cdist)
+
+  combinedReducer = (ee.Reducer.median().unweighted()
+                      .forEachBand(bandsOut.select(['SR_B1', 'SR_B2', 'SR_B3',
+                      'SR_B4', 'SR_B5', 'SR_B7']))
+    .combine(ee.Reducer.max().unweighted()
+      .forEachBand(bandsOut.select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 'SR_ATMOS_OPACITY'])), sharedInputs = False)
+    .combine(ee.Reducer.min().unweighted()
+      .forEachBand(bandsOut.select('ST_CDIST')), sharedInputs = False)
+    )
+  # Collect median reflectance and occurance values
+  # Make a cloud score, and get the water pixel count
+  data = (bandsOut
     .reduceRegions(
       collection = dt_label,
-      reducer = ee.Reducer.median().forEachBand(one_image),
+      reducer = combinedReducer,
       scale = 30,
-      tileScale = 2,
-      crs = one_image.geometry().projection().crs()
-      ))
-  
+      crs = one_image_SR_band.geometry().projection().crs()))
+      
   image_date_export = (ee.batch.Export.table.toDrive(
-    collection = data,
-    description = 'LS5_' + one_date.getInfo(),
-    folder = 'eePlumB_additional_band_data',
-    fileNamePrefix = 'LS5_' + one_date.getInfo() + '_additional_vars',
-    fileFormat = 'csv'))
+      collection = data,
+      description = 'LS5_' + one_date.getInfo(),
+      folder = 'eePlumB_additional_band_data',
+      fileNamePrefix = 'LS5_' + one_date.getInfo() + '_additional_vars_v2024-04-25',
+      fileFormat = 'csv'))
     
   image_date_export.start()
 

eePlumB/2_data_ingestion/3_data_re_pull_LS7.Rmd

@@ -89,7 +89,7 @@ export_l7_meta = (ee.batch.Export.table.toDrive(
   fileNamePrefix = 'LS7_image_metadata',
   fileFormat = 'csv'))
   
-export_l7_meta.start()
+#export_l7_meta.start()
 
 ```
 
@@ -99,6 +99,7 @@ And then apply the scaling factors to the stack
 l7 = (l7
   .map(rp.applyScaleFactors)
   .map(rp.apply_radsat_mask)
+  .map(rp.flag_qa_conf)
   .map(rp.addImageDate))
 
 ```
@@ -112,23 +113,47 @@ for i in range(dates_7.length().getInfo()):
   print(one_date.getInfo())
   one_dt = ee.Date(one_date)
   dt_label = labels_7_dt.filterDate(one_dt, one_dt.advance(1, 'day'))
-  one_image = l7.filterDate(one_dt, one_dt.advance(1, 'day')).mean()
-  
+  one_image_SR_band = (l7
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['SR_B1', 'SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B7'])
+    .mean())
+  one_image_bits = (l7
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 
+                          'SR_ATMOS_OPACITY'])
+    .max())
+  one_image_cdist = (l7
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select('ST_CDIST')
+    .min())
+
   #define bands to extract and reduce regions
-  data = (one_image
+  bandsOut = one_image_SR_band.addBands(one_image_bits).addBands(one_image_cdist)
+
+  combinedReducer = (ee.Reducer.median().unweighted()
+                      .forEachBand(bandsOut.select(['SR_B1', 'SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B7']))
+    .combine(ee.Reducer.max().unweighted()
+      .forEachBand(bandsOut.select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 'SR_ATMOS_OPACITY'])), sharedInputs = False)
+    .combine(ee.Reducer.min().unweighted()
+      .forEachBand(bandsOut.select('ST_CDIST')), sharedInputs = False)
+    )
+    
+  # Collect median reflectance and occurance values
+  # Make a cloud score, and get the water pixel count
+  data = (bandsOut
     .reduceRegions(
       collection = dt_label,
-      reducer = ee.Reducer.median().forEachBand(one_image),
+      reducer = combinedReducer,
       scale = 30,
-      tileScale = 2,
-      crs = one_image.geometry().projection().crs()
-      ))
+      crs = one_image_SR_band.geometry().projection().crs()))
   
   image_date_export = (ee.batch.Export.table.toDrive(
     collection = data,
     description = 'LS7_' + one_date.getInfo(),
     folder = 'eePlumB_additional_band_data',
-    fileNamePrefix = 'LS7_' + one_date.getInfo() + '_additional_vars',
+    fileNamePrefix = 'LS7_' + one_date.getInfo() + '_additional_vars_v2024-04-25',
     fileFormat = 'csv'))
     
   image_date_export.start()

eePlumB/2_data_ingestion/4_data_re_pull_LS8.Rmd

@@ -7,7 +7,9 @@ output: html_document
 
 # Purpose
 
-This script uses the volunteer point locations exported from `1_data_download.Rmd` which have been manually uploaded as a feature collection to the `ee-ross-superior` project in Earth Engine.
+This script uses the volunteer point locations exported from
+`1_data_download.Rmd` which have been manually uploaded as a feature collection
+to the `ee-ross-superior` project in Earth Engine.
 
 ## R/Python Setup
 
@@ -28,7 +30,8 @@ package_loader <- function(x) {
 invisible(lapply(libs, package_loader))
 ```
 
-Use the conda environment if the env folder is present, otherwise create the environment.
+Use the conda environment if the env folder is present, otherwise create the
+environment.
 
 ```{r conda-env}
 if (!dir.exists("env")) {
@@ -60,7 +63,8 @@ labels = ee.FeatureCollection("projects/ee-ross-superior/assets/labels/collated_
 labels_8 = labels.filter(ee.Filter.eq("mission", "LS8"))
 ```
 
-Get the unique dates of images from this list and assign the date to the `system:time_start` parameter.
+Get the unique dates of images from this list and assign the date to the
+`system:time_start` parameter.
 
 ```{python}
 dates_8 = labels_8.aggregate_array("date").distinct()
@@ -89,18 +93,21 @@ export_l8_meta = (ee.batch.Export.table.toDrive(
   fileNamePrefix = 'LS8_image_metadata',
   fileFormat = 'csv'))
   
-export_l8_meta.start()
+# export_l8_meta.start()
 
 ```
 
 And then apply the scaling factors to the stack
 
+And then apply the scaling factors to the stack
+
 ```{python}
 l8 = (l8
-  .map(rp.applyScaleFactors89)
+  .map(rp.applyScaleFactors_89)
   .map(rp.apply_radsat_mask)
-  .map(rp.addImageDate)
-  .map(rp.flag_cirrus_conf))
+  .map(rp.flag_qa_conf)
+  .map(rp.flag_high_aerosol)
+  .map(rp.addImageDate))
 
 ```
 
@@ -113,23 +120,47 @@ for i in range(dates_8.length().getInfo()):
   print(one_date.getInfo())
   one_dt = ee.Date(one_date)
   dt_label = labels_8_dt.filterDate(one_dt, one_dt.advance(1, 'day'))
-  one_image = l8.filterDate(one_dt, one_dt.advance(1, 'day')).mean()
+  one_image_SR_band = (l8
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['SR_B1', 'SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7'])
+    .mean())
+  one_image_bits = (l8
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 'aero_level'])
+    .max())
+  one_image_cdist = (l8
+    .filterDate(one_dt, one_dt.advance(1, 'day'))
+    .select('ST_CDIST')
+    .min())
+                        
   
   #define bands to extract and reduce regions
-  data = (one_image
+  bandsOut = one_image_SR_band.addBands(one_image_bits).addBands(one_image_cdist)
+
+  combinedReducer = (ee.Reducer.median().unweighted()
+                      .forEachBand(bandsOut.select(['SR_B1', 'SR_B2', 'SR_B3',
+                      'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7']))
+    .combine(ee.Reducer.max().unweighted()
+      .forEachBand(bandsOut.select(['cirrus_conf', 'snowice_conf', 'cloudshad_conf',
+                          'cloud_conf', 'dialated_cloud', 'aero_level'])), sharedInputs = False)
+    .combine(ee.Reducer.min().unweighted()
+      .forEachBand(bandsOut.select('ST_CDIST')), sharedInputs = False)
+    )
+  # Collect median reflectance and occurance values
+  # Make a cloud score, and get the water pixel count
+  data = (bandsOut
     .reduceRegions(
       collection = dt_label,
-      reducer = ee.Reducer.median().forEachBand(one_image),
+      reducer = combinedReducer,
       scale = 30,
-      tileScale = 2,
-      crs = one_image.geometry().projection().crs()
-      ))
-  
+      crs = one_image_SR_band.geometry().projection().crs()))
+      
   image_date_export = (ee.batch.Export.table.toDrive(
     collection = data,
     description = 'LS8_' + one_date.getInfo(),
     folder = 'eePlumB_additional_band_data',
-    fileNamePrefix = 'LS8_' + one_date.getInfo() + '_additional_vars',
+    fileNamePrefix = 'LS8_' + one_date.getInfo() + '_additional_vars_v2024-04-25',
     fileFormat = 'csv'))
     
   image_date_export.start()