diff --git a/vignettes/ars.Rmd b/vignettes/ars.Rmd
deleted file mode 100644
index cfc2a2b..0000000
--- a/vignettes/ars.Rmd
+++ /dev/null
@@ -1,130 +0,0 @@
----
-title: "ARS - Spectral Unmixing"
-author: "Mike Cecil"
-date: "2023-02-26"
-output: html_document
----
-
-```{r setup, include=FALSE}
-knitr::opts_chunk$set(echo = TRUE)
-```
-
-Install `RStoolbox` from source. 
-
-```{r}
-library(devtools)
-install_github("bleutner/RStoolbox")
-```
-
-This may throw an error, in which case install any other dependencies then rerun above. 
-
-```{r}
-#install.packages("", "", "") ## install dependencies
-```
-
-Load other libraries
-```{r cars}
-library(dplyr)
-library(raster)
-library(sf)
-library(RStoolbox)
-```
-
-Set up file paths 
-```{r}
-taxonomic_path <- "C:/Users/micha/Downloads/Endmembers/TaxonomicEndmembers.shp"
-biophysical_path <- "C:/Users/micha/Downloads/Endmembers/BiophysicalEndmembers.shp"
-landsat_path <- "C:/Users/micha/Downloads/Landsat8SR_SD_July2019_scale.tif"
-```
-
-Read in data
-```{r}
-tax_classes <- st_read (taxonomic_path)
-#bio_classes <- st_read(biophysical_path)
-lsat <- raster::brick(landsat_path)
-```
-
-Plot data
-```{r}
-plot(lsat[[4]])
-plot(tax_classes, add = T)
-```
-Extract pixel values at point locations.
-
-```{r}
-tax_em <- extract(lsat, tax_classes %>% st_centroid()) %>% 
-  cbind(tax_classes, .) %>% as_tibble()
-tax_em
-```
-Because we have two points per land cover class, we will use `group_by` to average
-```{r}
-tax_group <- tax_em %>% group_by(class) %>%
-  summarise(mean_b1 = mean(Landsat8SR_SD_July2019_scale_1),
-            mean_b2 = mean(Landsat8SR_SD_July2019_scale_2),
-            mean_b3 = mean(Landsat8SR_SD_July2019_scale_3),
-            mean_b4 = mean(Landsat8SR_SD_July2019_scale_4),
-            mean_b5 = mean(Landsat8SR_SD_July2019_scale_5)) %>%
-  st_drop_geometry()
-tax_group
-```
-
-We need our spectral table to only have a column for each band. So we'll put the land cover classes into row names (instead of as a column.)
-
-```{r}
-
-class_names <-  tax_group$class
-
-tax_group <- tax_group %>% data.frame() %>%  dplyr::select(-class)
-rownames(tax_group) <- class_names
-tax_group
-```
-
-Perform spectral unmixing. The issue with this algorithm is that the class percents may not add to 1. (this may take 10-20 min)
-
-```{r}
-
-probs_tax <- mesma(lsat,
-                 tax_group,
-                 method = "NNLS",
-                 iterate = 1000)
-
-plot(probs_tax)
-```
-
-The code below is used to "rescale" the probabilities so they add to 1. 
-
-```{r}
-sum_tax <- sum(probs_tax) - probs_tax[['RMSE']]
-probs_tax_adj <- probs_tax
-## for loop rescales each band
-for(band in class_names){
- # print(band)
-  probs_tax_adj[[band]] <- probs_tax_adj[[band]]/sum_tax
-}
-
-summary(probs_tax_adj)
-
-
-```
-
-Let's check that the adjusted values sum to 1.
-
-```{r}
-sum_tax_adj <- sum(probs_tax_adj) - probs_tax_adj[['RMSE']] ## sum all bands except RMSE
-summary(sum_tax_adj)
-plot(sum_tax_adj)
-```
-
-Now let's plot the final class percents
-
-```{r}
-plot(probs_tax_adj, zlim = c(0,1))
-```
-
-And we can save the results to a tif file.
-
-```{r}
-output_folder <- "C:/Users/micha/Downloads/"
-writeRaster(probs_tax_adj, file = paste0(output_folder, "class_probabilities.tif"))
-```
-