Updates authors, adds figure captions, clarifies specimen organization

README.md

@@ -3,8 +3,7 @@ license: cc-by-4.0
 language:
 - en
 - la
-pretty_name: 'Beetles as Sentinel Taxa: Predicting drought conditions from NEON specimen
-  imagery'
+pretty_name: "Beetles as Sentinel Taxa: Predicting drought conditions from NEON specimen imagery"
 task_categories:
 - image-feature-extraction
 dataset_info:
@@ -72,6 +71,7 @@ configs:
     path: data/validation-*
 ---
 
+
 <!--
 Image with caption (jpg or png):
 |![Figure #](https://huggingface.co/datasets/imageomics/<data-repo>/resolve/main/<filepath>)|
@@ -106,19 +106,19 @@ This dataset contains images of pinned carabid beetle specimens collected by the
     - David Carlyn, The Ohio State University, Columbus, OH, USA
     - Connor Kilrain, The Ohio State University, Columbus, OH, USA
     - Jiaman Wu, The Ohio State University, Columbus, OH, USA
-    - Elizabeth G. Campolongo, The Ohio State University, Columbus, OH, USA
-    - Wei-Lun Chao, The Ohio State University, Columbus, OH, USA
     - Chandra Earl, National Ecological Observatory Network (NEON), Tempe, AZ, USA
     - Hilmar Lapp, Duke University, Durham, NC, USA
     - Kayla I. Perry, The Ohio State University, Wooster, OH, USA
-    - Sydne Record, University of Maine, Orono, ME, USA
-    - Eric R. Sokol, National Ecological Observatory Network (NEON), Boulder, CO, USA  
     - Matthew J. Thompson, The Ohio State University, Columbus, OH, USA
+    - Elizabeth G. Campolongo, The Ohio State University, Columbus, OH, USA
+    - Wei-Lun Chao, The Ohio State University, Columbus, OH, USA
+    - Eric R. Sokol, National Ecological Observatory Network (NEON), Boulder, CO, USA
+    - Sydne Record, University of Maine, Orono, ME, USA
 - **Homepage:** https://www.nsfhdr.org/mlchallenge-y2
 - **Repository:** https://github.com/Imageomics/HDR-SMood-Challenge
 <!-- - **Paper:**  TBD -->
 
-This dataset contains high-resolution images and metadata for pinned specimens of carabid beetles collected across U.S. ecosystems by the [National Ecological Observatory Network (NEON)](https://www.neonscience.org/). Each image is linked to specimen-level metadata (e.g., collection date, site, taxonomic identification) and environmental context, including drought severity indicators ([Standardized Precipitation Evapotranspiration Index (SPEI)](https://spei.csic.es/)) calculated over multiple timescales using remote sensing data. This dataset is designed to support research on the relationship between ecological traits and climate stress, and is intended for training and evaluating machine learning models that predict environmental conditions-particularly drought status—from biological imagery. It was created for the Imageomics portion of the [second HDR ML Challenge](https://www.nsfhdr.org/mlchallenge-y2), which emphasizes model generalization across sites with different climates, land cover types, and beetle species pools. Not all available information is provided as part of the challenge, but it will be added at the end of the challenge to allow for broader use beyond the challenge.
+This dataset contains high-resolution images and metadata for pinned specimens of carabid beetles collected across U.S. ecosystems by the [National Ecological Observatory Network (NEON)](https://www.neonscience.org/). Each image is linked to specimen-level metadata (e.g., collection date, site, taxonomic identification) and environmental context, including drought severity indicators ([Standardized Precipitation Evapotranspiration Index (SPEI)](https://spei.csic.es/)) calculated over multiple timescales using remote sensing data. This dataset is designed to support research on the relationship between ecological traits and climate stress, and is intended for training and evaluating machine learning models that predict environmental conditions-particularly drought status—from biological imagery. It was created for the Imageomics portion of the [second HDR ML Challenge](https://www.nsfhdr.org/mlchallenge-y2), which emphasizes model generalization across sites with different climates, land cover types, and beetle species pools. Not all available information is provided as part of the challenge, but it will be added at the end of the challenge to allow for broader use beyond the challenge. This includes smaller beetles, which were entirely excluded from the challenge, but will be added to this dataset later.
 
 ### Supported Tasks and Leaderboards
 
@@ -150,7 +150,7 @@ Leaderboard is available on the [Codabench Challenge page](https://www.codabench
 
 ### Data Instances
 
-Each record in `train.csv` or `val.csv` corresponds to a single pinned carabid beetle specimen collected by [NEON](https://www.neonscience.org/) staff as part of the "Ground beetles sampled from pitfall traps" data product ([DP1.10022.001](https://data.neonscience.org/data-products/DP1.10022.001)). There are a maximum of 13 field season collection bouts per year, with carabids collected from no more than 10 plotIDs per bout. For data collected prior to 2018, each plot will yield no more than 4 samples per bout of collection, resulting in a maximum of 520 plot‐bouts per site per year. For collections 2018 and later, each plot yields a maximum of 3 samples within each bout. The number of individuals identified varies with the abundance of organisms at the site. Beetles were imaged in the trays in which they were pinned, which correspond to the taxonomic designation but may not align with the collection date; individuals were cropped from the group images, and saved in `flatten_images/`. As a result of this imaging process, each color palette and scalebar image corresponds to more than one beetle image (alignment is described below).
+Each record in `train.csv` or `val.csv` corresponds to a single pinned carabid beetle specimen collected by [NEON](https://www.neonscience.org/) staff as part of the "Ground beetles sampled from pitfall traps" data product ([DP1.10022.001](https://data.neonscience.org/data-products/DP1.10022.001)). There are a maximum of 13 field season collection bouts per year, with carabids collected from no more than 10 plotIDs per bout. For data collected prior to 2018, each plot will yield no more than 4 samples per bout of collection, resulting in a maximum of 520 plot‐bouts per site per year. For collections 2018 and later, each plot yields a maximum of 3 samples within each bout. The number of individuals identified varies with the abundance of organisms at the site. Beetles were imaged in the trays in which they were pinned, which correspond to their taxonomic designations; individuals were cropped from the group images, and saved in `flatten_images/`. As a result of this imaging process, each color palette and scalebar image corresponds to more than one beetle image (alignment is described below).
 
 Metadata include an identifier for a collection event (`eventID`), the date of the collection event (`collectDate`), an anonymized identifier of the domain (`domainID`) and site (`siteID`) where the collection event took place, taxonomic information (`scientificName`), unique beetle image identifier (`public_id`), a link to the beetle image file (`relative_img_loc`), a link to the color palette image (`colorpicker_path`), a link to the scale image (`scalebar_path`), and Standardized Precipication Evapotranspiratoin Index (SPEI) values that correspond with the location and time that the specimen was collected. The SPEI values were calculated for the 30 day (`SPEI_30d`), 1 year (`SPEI_1y`), and 2 year (`SPEI_2y`) time windows preceding the time of collection at each location for each beetle specimen in the dataset. 
 
@@ -178,8 +178,9 @@ Both `train.csv` and `val.csv` have the following columns.
 
 ### Data Splits
 
-![fictional map to visualize the data splits as described below (there are 1-3 sites per domain, where each domain represents a region that is reasonably climactically consistent)](https://huggingface.co/datasets/imageomics/sentinel-beetles/resolve/main/figures/fictional_map.png)
-_Image was created with Microsoft CoPilot + manual editing_
+| ![fictional map to visualize the data splits as described below (there are 1-3 sites per domain, where each domain represents a region that is reasonably climactically consistent)](https://huggingface.co/datasets/imageomics/sentinel-beetles/resolve/main/figures/fictional_map.png)|
+|:--|
+|**Figure 1.** Fictional depiction of data distribution and splits across domains. _Image was created with Microsoft CoPilot + manual editing_.|
 
 Using the image above as a guide, the data has been split in the following way:
 
@@ -231,12 +232,13 @@ This dataset consists of images of _pinned_ carabid beetle specimens taken in 20
 #### Data Collection and Processing
 
 **Specimen collection:**  
-NEON beetle specimens were collected using pitfall traps—16 oz deli containers filled with 150 mL or 250 mL of propylene glycol. Three traps were deployed at 10 distributed base plots per site for two-week periods throughout the growing season. Traps were positioned at the East, West, and South edges of the plots (20 meters from center). After field collection, arthropod samples were sorted in the lab. Carabid beetles were identified to species where possible. A subset of these were pinned or pointed, and from those, some were forwarded for taxonomic review or DNA barcoding. Beetles are pinned in trays, and organized into these trays based on the taxonomic labeling (i.e., "expert" or "parataxonomist", described in [Annotations](#annotations) below), thus, inclusion in the same tray is only indicative of matching taxonomy.
+NEON beetle specimens were collected using pitfall traps—16 oz deli containers filled with 150 mL or 250 mL of propylene glycol. Three traps were deployed at 10 distributed base plots per site for two-week periods throughout the growing season. Traps were positioned at the East, West, and South edges of the plots (20 meters from center). After field collection, arthropod samples were sorted in the lab. Carabid beetles were identified to species where possible. A subset of these were pinned or pointed, and from those, some were forwarded for taxonomic review or DNA barcoding. Beetles are pinned in trays, and organized into these trays based on the taxonomic labeling, species identification, domain ID, and collection year. There are known instances of errors in tray organization where there are multiple species, years. or domains in one tray. Thus, inclusion in the same tray is usually, but not always indicative of matching taxonomy, collection year, or domain. Theses errors are corrected in the individual metadata records.
 
 **Imaging workflow:** 
 
-![fictional drawer containing beetles of the same species, a color calibration card, and a scale. The beetles may be collected from separate events.](https://huggingface.co/datasets/imageomics/sentinel-beetles/resolve/main/figures/fictional_beetle_tray.png)
-_Image was created with Microsoft CoPilot + manual editing_
+|![fictional drawer containing beetles of the same species, a color calibration card, and a scale. The beetles may be collected from separate events.](https://huggingface.co/datasets/imageomics/sentinel-beetles/resolve/main/figures/fictional_beetle_tray.png)|
+|:--|
+|**Figure 2.** Artificial depiction of a drawer containing one species of beetles with a color calibration card and a scalebar. _Image was created with Microsoft CoPilot + manual editing_. |
 
 
 Beetles were imaged in bulk (by tray) with a color palette and scalebar included in the image for later standardization. As noted above, beetles from different events may be in the same trays and same event in different trays based on labeling, so use the `colorpicker_path` and `scalbar_path` indicators to ensure proper alignment. The scalebars were placed at the same height as the pinned specimens in the tray to ensure consistent distance from the camera.
@@ -313,8 +315,8 @@ No warranties are given. The license may not give you all of the permissions nec
 ```
 @misc{<ref_code>,
   author = {Alyson East and Michael Belitz and Leah Cotton and Jacqueline Dominguez and Isabelle Betancourt and 
-    S M Rayeed and Fangxun Liu and David Carlyn and Connor Kilrain and Jiaman Wu and Elizabeth G. Campolongo and 
-    Wei-Lun Chao and Chandra Earl and Hilmar Lapp and Kayla I. Perry and Sydne Record and Eric R. Sokol and Matthew J. Thompson},
+    S M Rayeed and Fangxun Liu and David Carlyn and Connor Kilrain and Jiaman Wu and Chandra Earl and Hilmar Lapp 
+    and Kayla I. Perry and Matthew J. Thompson and Elizabeth G. Campolongo and Wei-Lun Chao and Eric R. Sokol and Sydne Record},
   title = {Beetles as Sentinel Taxa: Predicting drought conditions from NEON specimen imagery},
   year = {2025},
   url = {https://huggingface.co/datasets/imageomics/sentinel-beetles},
@@ -371,4 +373,4 @@ The HDR ML Challenge program is hosting its second FAIR challenge, this year pre
 
 ## Dataset Card Contact
 
-We encourage those with questions regarding the dataset to open a discussion in the [Community tab](https://huggingface.co/datasets/imageomics/sentinel-beetles/discussions).
+We encourage those with questions regarding the dataset to open a discussion in the [Community tab](https://huggingface.co/datasets/imageomics/sentinel-beetles/discussions).