README.md

---
license: cc-by-4.0
language:
- en
pretty_name: Hawaii PUUM Individual Phenology Dataset
task_categories: 
- time-series-forecasting # ex: image-classification, see key list at https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/src/pipelines.ts
tags:
- biology
- image
- CV
- time series
- phenology
- greenness
- clustering
- discrete wavelet transform
- hawaii
- O'hia
- koa
- tree
size_categories: n<1K # ex: n<1K, 1K<n<10K, 10K<n<100K, 100K<n<1M, ...
description: >-
  This dataset extracts greenness and redness curves of individual O'hia and Koa
  trees in the Pu'u Maka'ala Natural Area Reserve (PUUM) on the island of
  Hawaii (Big Island). The dataset contains raw, 1-day, and 3-day summaries of statistics.
---

<!--

NOTE: Add more tags (your particular animal, type of model and use-case, etc.).

As with your GitHub Project repo, it is important to choose an appropriate license for your dataset. The default license is [CC0](https://creativecommons.org/publicdomain/zero/1.0/) (public domain dedication, see [Dryad's explanation of why to use CC0](https://blog.datadryad.org/2023/05/30/good-data-practices-removing-barriers-to-data-reuse-with-cc0-licensing/)). Alongside the appropriate stakeholders (eg., your PI, co-authors), select a license that is [Open Source Initiative](https://opensource.org/licenses) (OSI) compliant.
For more information on how to choose a license and why it matters, see [Choose A License](https://choosealicense.com) and [A Quick Guide to Software Licensing for the Scientist-Programmer](https://doi.org/10.1371/journal.pcbi.1002598) by A. Morin, et al.
See the [Imageomics policy for licensing](https://imageomics.github.io/Imageomics-guide/wiki-guide/Digital-products-release-licensing-policy/) for more information.

See more options for the above information by clicking "edit dataset card" on your repo.

Fill in as much information as you can at each location that says "More information needed".
-->

<!--
Image with caption (jpg or png):
|![Figure #](https://huggingface.co/datasets/imageomics/<data-repo>/resolve/main/<filepath>)|
|:--|
|**Figure #.** [Image of <>](https://huggingface.co/datasets/imageomics/<data-repo>/raw/main/<filepath>) <caption description>.|
-->

<!--
Notes on styling:

To render LaTex in your README, wrap the code in `\\(` and `\\)`. Example: \\(\frac{1}{2}\\)

Escape underscores ("_") with a "\". Example: image\_RGB
-->

# Dataset Card for Phenology-Normal-Hawaii

<!-- Provide a quick summary of what the dataset is or can be used for. --> 

## Dataset Details

### Dataset Description

- **Curated by:** Jianyang Gu, Faye Xie, Colby Stakun-Pickering
<!-- Provide the basic links for the dataset. These will show up on the sidebar to the right of your dataset card ("Curated by" too). -->
<!-- - **Homepage:**  -->
- **Repository:** [phenology-project-HI](https://github.com/Imageomics/phenology-project-HI#)
<!-- - **Paper:**  -->


<!-- Provide a longer summary of what this dataset is. -->
This dataset provides time series of vegetation color indices collected from the PUUM ([Pu'u Maka'ala](https://www.neonscience.org/field-sites/puum)) site, designed to support fine-grained phenological analysis. The time series are extracted from the [PhenoCam](https://phenocam.nau.edu/webcam/) images captured by [NEON](https://www.neonscience.org/) at the PUUM site. It includes three types of GCC ([Green Chromatic Coordinate](https://phenocam.nau.edu/education/U3_PhenoCamData_Activity_Matching.pdf)) and RCC (Red Chromatic Coordinate) curves:
* Raw curves: Original time series extracted from masked individual crowns without additional processing.
* Brightest-pixel curves: Curves filtered using K-means clustering to retain only the brightest pixels, helping to reduce noise from shadows and background variation (specifically for Koa trees).
* Smoothed curves: Curves further processed with a discrete wavelet transform to denoise and smooth short-term fluctuations, preserving major seasonal trends.

All curve data are stored in CSV files, with each file containing a time series for a specific individual and curve type. The CSV format ensures easy loading and integration into common data analysis workflows.

Each curve tracks vegetation color changes at the level of individual tree crowns across multiple seasons. The dataset supports research into individual phenology patterns, intra-site variability, and ecological modeling that requires high-temporal-resolution vegetation signals.

The three curve types allow flexible usage depending on the desired balance between fidelity to raw data and robustness to noise.

<!--This dataset card aims to be a base template for new datasets. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1), and further altered to suit Imageomics Institute needs.-->


### Supported Tasks
The dataset supports time-series forecasting.

<!-- Provide benchmarking results -->


## Dataset Structure

```
/dataset/
    timeseries/
        EB_0001/
            RAW/
                NEON.D20.PUUM.DP1.00033_EB_0001_roistats.csv
                NEON.D20.PUUM.DP1.00033_EB_0001_1day.csv
                NEON.D20.PUUM.DP1.00033_EB_0001_3day.csv
            KMEANS/
                NEON.D20.PUUM.DP1.00033_EB_0001_roistats.csv
                NEON.D20.PUUM.DP1.00033_EB_0001_1day.csv
                NEON.D20.PUUM.DP1.00033_EB_0001_3day.csv
            DWT/
                NEON.D20.PUUM_EB_0001_trend_summary.csv
        ...
        EB_0004/
            RAW/
                NEON.D20.PUUM.DP1.00033_EB_0004_roistats.csv
                NEON.D20.PUUM.DP1.00033_EB_0004_1day.csv
                NEON.D20.PUUM.DP1.00033_EB_0004_3day.csv
            DWT/
                NEON.D20.PUUM_EB_0001_trend_summary.csv
        ...
```

`NEON.D20.PUUM.DP1.00033` indicates the site and camera index. `EB_0001` means the generated instance mask. 
The first three masks are for Koa trees, where greenness is our main focus. Therefore, k-means is applied to extract more subtle signals.
`EB_0004` and `EB_0005` are for O'hia trees, where we want to extract red flowers. K-means is not applied as the signal is more distinguishable. 

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->

<!-- Provide format of the dataset, ex:

```​
/dataset/
    <species_1>/
        <img_id 1>.png
        <img_id 2>.png
        ...
        <img_id n>.png
    <species_2>/
        <img_id 1>.png
        <img_id 2>.png
        ...
        <img_id n>.png
    ...
    <species_N>/
        <img_id 1>.png
        <img_id 2>.png
        ...
        <img_id n>.png
    metadata.csv
```​

-->

### Data Instances
All the files are time-series of GCC (some also include RCC), under the folder of corresponding individual masks. The `RAW` folder contains time-series acquired by simply applying the basic mask. The `KMEANS` folder contains time-series calculated with pixels of high brightness. And the `DWT` folder contains time-series after discrete wavelet transform. 

<!--
Describe data files

Ex: All images are named <img_id>.png, each within a folder named for the species. They are 1024 x 1024, and the color has been standardized using <link to color standardization package>.
-->

### Data Fields

**roistats.csv**: raw statistics of the region of interet. 
| Column Name   | Description |
|:--------------|:------------|
| `date` | Date when the image was taken (YYYY-MM-DD) |
| `local_std_time` | Local standard time when the image was captured |
| `doy` | Day of year (1–365/366) corresponding to the image date |
| `filename` | Name of the image file |
| `solar_elev` | Solar elevation angle at the time of image capture (degrees) |
| `exposure` | Exposure setting used for the image |
| `awbflag` | Automatic white balance flag indicating image color adjustment status |
| `mask_index` | Index identifying the masked region of interest within the image |
| `gcc` | Green chromatic coordinate (proportion of green in total brightness) |
| `rcc` | Red chromatic coordinate (proportion of red in total brightness) |
| `r_mean` | Mean red channel value of selected pixels |
| `r_std` | Standard deviation of red channel values |
| `r_5_qtl` | 5th percentile of red channel values |
| `r_10_qtl` | 10th percentile of red channel values |
| `r_25_qtl` | 25th percentile (first quartile) of red channel values |
| `r_50_qtl` | 50th percentile (median) of red channel values |
| `r_75_qtl` | 75th percentile (third quartile) of red channel values |
| `r_90_qtl` | 90th percentile of red channel values |
| `r_95_qtl` | 95th percentile of red channel values |
| `g_mean` | Mean green channel value of selected pixels |
| `g_std` | Standard deviation of green channel values |
| `g_5_qtl` | 5th percentile of green channel values |
| `g_10_qtl` | 10th percentile of green channel values |
| `g_25_qtl` | 25th percentile (first quartile) of green channel values |
| `g_50_qtl` | 50th percentile (median) of green channel values |
| `g_75_qtl` | 75th percentile (third quartile) of green channel values |
| `g_90_qtl` | 90th percentile of green channel values |
| `g_95_qtl` | 95th percentile of green channel values |
| `b_mean` | Mean blue channel value of selected pixels |
| `b_std` | Standard deviation of blue channel values |
| `b_5_qtl` | 5th percentile of blue channel values |
| `b_10_qtl` | 10th percentile of blue channel values |
| `b_25_qtl` | 25th percentile (first quartile) of blue channel values |
| `b_50_qtl` | 50th percentile (median) of blue channel values |
| `b_75_qtl` | 75th percentile (third quartile) of blue channel values |
| `b_90_qtl` | 90th percentile of blue channel values |
| `b_95_qtl` | 95th percentile of blue channel values |
| `r_g_correl` | Correlation coefficient between red and green channel values |
| `g_b_correl` | Correlation coefficient between green and blue channel values |
| `b_r_correl` | Correlation coefficient between blue and red channel values |


**1day/3day.csv**: 1-day/3-day stats over the GCC and RCC.
| Column Name | Description |
|:------------|:------------|
| `date` | Date of observation (YYYY-MM-DD) |
| `year` | Year of observation |
| `doy` | Day of year (1–365/366) corresponding to the observation date |
| `image_count` | Number of images taken on that date |
| `midday_filename` | Filename of the midday image |
| `midday_r` | Mean red channel value of the midday image |
| `midday_g` | Mean green channel value of the midday image |
| `midday_b` | Mean blue channel value of the midday image |
| `midday_gcc` | Green chromatic coordinate (GCC) calculated from the midday image |
| `midday_rcc` | Red chromatic coordinate (RCC) calculated from the midday image |
| `r_mean` | Mean red channel value across all daily images |
| `r_std` | Standard deviation of red channel values across daily images |
| `g_mean` | Mean green channel value across all daily images |
| `g_std` | Standard deviation of green channel values across daily images |
| `b_mean` | Mean blue channel value across all daily images |
| `b_std` | Standard deviation of blue channel values across daily images |
| `gcc_mean` | Mean green chromatic coordinate (GCC) across daily images |
| `gcc_std` | Standard deviation of GCC values across daily images |
| `gcc_50` | 50th percentile (median) of GCC values |
| `gcc_75` | 75th percentile of GCC values |
| `gcc_90` | 90th percentile of GCC values |
| `rcc_mean` | Mean red chromatic coordinate (RCC) across daily images |
| `rcc_std` | Standard deviation of RCC values across daily images |
| `rcc_50` | 50th percentile (median) of RCC values |
| `rcc_75` | 75th percentile of RCC values |
| `rcc_90` | 90th percentile of RCC values |
| `max_solar_elev` | Maximum solar elevation angle recorded during the day |
| `snow_flag` | Flag indicating presence of snow (1 if snow detected, 0 otherwise) |
| `outlierflag_gcc_mean` | Outlier flag for the daily mean GCC value |
| `outlierflag_gcc_50` | Outlier flag for the 50th percentile GCC value |
| `outlierflag_gcc_75` | Outlier flag for the 75th percentile GCC value |
| `outlierflag_gcc_90` | Outlier flag for the 90th percentile GCC value |

**trend_summary.csv**: smoothed GCC and RCC.
| Column Name | Description |
|:------------|:------------|
| `img_date` | Date of observation (YYYY-MM-DD) |
| `A4D4` | The 4th-level approximation and detail coefficients obtained from the discrete wavelet transform |
| `A8D8` | The 8th-level approximation and detail coefficients obtained from the discrete wavelet transform |

<!--
Describe the types of the data files or the columns in a CSV with metadata.

Ex: 
**metadata.csv**:
  - `img_id`: Unique identifier for the dataset. 
  - `specimen_id`: ID of specimen in the image, provided by museum data source. There are multiple images of a single specimen.
  - `species`: Species of the specimen in the image. There are N different species of <genus> of <animal>.
  - `view`: View of the specimen in the image (e.g., `ventral` or `dorsal` OR `top` or `bottom`, etc.; specify options where reasonable).
  - `file_name`: Relative path to image from the root of the directory (`<species>/<img_id>.png`); allows for image to be displayed in the dataset viewer alongside its associated metadata.
-->

<!-- ### Data Splits
[More Information Needed] -->
<!--
Give your train-test splits for benchmarking; could be as simple as "split is indicated by the `split` column in the metadata file: `train`, `val`, or `test`." Or perhaps this is just the training dataset and other datasets were used for testing (you may indicate which were used).
-->

## Dataset Creation

### Curation Rationale
The dataset contains individual GCC and RCC curves for the PUUM site. It was created to enable detailed analysis of vegetation phenology at the level of individual tree crowns, rather than relying solely on site-level averages. By extracting and tracking GCC (Green Chromatic Coordinate) and RCC (Red Chromatic Coordinate) values for specific regions over time, the dataset provides fine-grained insights into seasonal patterns, canopy development, and potential variability among individuals.
<!-- Motivation for the creation of this dataset. For instance, what you intended to study and why that required curation of a new dataset (or if it's newly collected data and why the data was collected (intended use)), etc. -->

### Source Data
[PhenoCam images](https://phenocam.nau.edu/webcam/) in [PUUM](https://phenocam.nau.edu/webcam/sites/NEON.D20.PUUM.DP1.00033/) site were downloaded. This data product is developed upon the proposed [GCC generation pipeline](https://github.com/Imageomics/phenology-project-HI/tree/main/gcc_generation) to generate phenological time-series based on the images. 

To produce this dataset, images are first gathered from [PUUM PhenoCam](https://phenocam.nau.edu/webcam/sites/NEON.D20.PUUM.DP1.00033/). We use the images from 10am to 2pm every day for best lighting conditions. Please create an account in the PhenoCam page, and download the image data between 10am to 2pm. To fully replicate our results, please download the data from 2019-04-28 to 2025-04-24. We then manually label individual masks for one example image (see [below](#annotation-process) for more details). The raw time-series are generated by directly applying the mask to the image. Then we conduct K-means clustering on the individual pixels of every image. As the flowers of Koa trees are of white color, we only keep the pixels with the largest brightness to perform GCC calculation. And finally, we apply discrete wavelet transform to get the smoothed time-series. 
<!-- This section describes the source data (e.g., news text and headlines, social media posts, translated sentences, ...). As well as an original source it was created from (e.g., sampling from Zenodo records, compiling images from different aggregators, etc.) -->

<!-- #### Data Collection and Processing -->
<!-- [More Information Needed] -->
<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, re-sizing of images, tools and libraries used, etc. 
This is what _you_ did to it following collection from the original source; it will be overall processing if you collected the data initially.
-->

<!-- #### Who are the source data producers? -->
<!-- [More Information Needed] -->
<!-- This section describes the people or systems who originally created the data.

Ex: This dataset is a collection of images taken of the butterfly collection housed at the Ohio State University Museum of Biological Diversity. The associated labels and metadata are the information provided with the collection from biologists that study butterflies and supplied the specimens to the museum.
 -->


### Annotations
We provide manual annotations of individual masks. Each mask identifies one tree individual. Based on the acquired masks, we can perform phenological analysis for different individuals. 
<!-- 
If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. 

Ex: We standardized the taxonomic labels provided by the various data sources to conform to a uniform 7-rank Linnean structure. (Then, under annotation process, describe how this was done: Our sources used different names for the same kingdom (both _Animalia_ and _Metazoa_), so we chose one for all (_Animalia_). -->

#### Annotation process
The annotation is conducted through an interactive interface created with [matplotlib](https://matplotlib.org/). Users are able to draw the contour for the individual of interest. Then the mask will be automatically saved based on the contour. The interface script is available in [our repository](https://github.com/Imageomics/phenology-project-HI/blob/main/gcc_generation/draw.py).
<!-- This section describes the annotation process such as annotation tools used, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->

#### Who are the annotators?
Faye Xie
<!-- This section describes the people or systems who created the annotations. -->

### Personal and Sensitive Information
N/A
<!-- 
For instance, if your data includes people or endangered species. -->


## Considerations for Using the Data
Please cite this dataset if you use it.
<!-- [More Information Needed] -->
<!--
Things to consider while working with the dataset. For instance, maybe there are hybrids and they are labeled in the `hybrid_stat` column, so to get a subset without hybrids, subset to all instances in the metadata file such that `hybrid_stat` is _not_ "hybrid".
-->

### Bias, Risks, and Limitations
The dataset is biased towards the individuals close to the PhenoCam. While the observed individuals within one species demonstrate similar phenological timing, the conclusion doesnt't apply to the other individuals or other sites in the island. 
<!-- This section is meant to convey both technical and sociotechnical limitations. Could also address misuse, malicious use, and uses that the dataset will not work well for.-->

<!-- For instance, if your data exhibits a long-tailed distribution (and why). -->

### Recommendations
We don't recommend using the conclusion drawn from this dataset to infer the other sites in Hawaii. 
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

## Licensing Information
This product is licensed under the [Creative Commons Attribution 4.0 International License (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/). This means the material can be freely shared, copied, redistributed, adapted, and built upon for any purpose, including commercial use, as long as appropriate credit is given to the original source. Users must provide attribution, indicate if changes were made, and link to the license. No additional restrictions beyond those stated in the license terms may be applied.

<!-- See notes at top of file about selecting a license. 
If you choose CC0: This dataset is dedicated to the public domain for the benefit of scientific pursuits. We ask that you cite the dataset and journal paper using the below citations if you make use of it in your research.

Be sure to note different licensing of images if they have a different license from the compilation.
ex: 
The data (images and text) contain a variety of licensing restrictions mostly within the CC family. Each image and text in this dataset is provided under the least restrictive terms allowed by its licensing requirements as provided to us (i.e, we impose no additional restrictions past those specified by licenses in the license file).

EOL images contain a variety of licenses ranging from [CC0](https://creativecommons.org/publicdomain/zero/1.0/) to [CC BY-NC-SA](https://creativecommons.org/licenses/by-nc-sa/4.0/).
For license and citation information by image, see our [license file](https://huggingface.co/datasets/imageomics/treeoflife-10m/blob/main/metadata/licenses.csv).

This dataset (the compilation) has been marked as dedicated to the public domain by applying the [CC0 Public Domain Waiver](https://creativecommons.org/publicdomain/zero/1.0/). However, images may be licensed under different terms (as noted above).
-->

## Citation

**BibTeX:**
```
@misc{phenology_normal_hawaii
    author = {Jianyang Gu and Faye Xie and Colby Stakun-Pickering and Adam Young and Chris Florian},
    title = {Hawaii PUUM Individual Phenology Dataset (Revision 7c66375)},
    year = {2025},
    url = {https://huggingface.co/datasets/imageomics/phenology_normal_hawaii},
    doi = {10.57967/hf/7304},
    publisher = {Hugging Face}
}
```

Please also cite the PhenoCam images:
```
@misc{DP1.00033.001/provisional,
  url = {https://data.neonscience.org/data-products/DP1.00033.001},
  author = {{National Ecological Observatory Network (NEON)}},
  language = {en},
  title = {Phenology images (DP1.00033.001)},
  publisher = {National Ecological Observatory Network (NEON)},
  year = {2025}
}
```

<!--
If you want to include BibTex, replace "<>"s with your info 

**Data**
```​
@misc{<ref_code>,
  author = {<author1 and author2>},
  title = {<title>},
  year = {<year>},
  url = {https://huggingface.co/datasets/imageomics/<dataset_name>},
  doi = {<doi once generated>},
  publisher = {Hugging Face}
}
```​

-for an associated paper:
**Paper**
```​
@article{<ref_code>,
  title    = {<title>},
  author   = {<author1 and author2>},
  journal  = {<journal_name>},
  year     =  <year>,
  url      = {<DOI_URL>},
  doi      = {<DOI>}
}
```​
-->

<!---
If the data is modified from another source, add the following. 

Please be sure to also cite the original data source(s):
<citation>
-->


## Acknowledgements

This work was supported by both the [Imageomics Institute](https://imageomics.org) and the [AI and Biodiversity Change (ABC) Global Center](http://abcresearchcenter.org). The Imageomics Institute is funded by the US National Science Foundation's Harnessing the Data Revolution (HDR) program under [Award #2118240](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2118240) (Imageomics: A New Frontier of Biological Information Powered by Knowledge-Guided Machine Learning). The ABC Global Center is funded by the US National Science Foundation under [Award No. 2330423](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2330423&HistoricalAwards=false) and Natural Sciences and Engineering Research Council of Canada under [Award No. 585136](https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=782440). This dataset draws on research supported by the Social Sciences and Humanities Research Council. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation, Natural Sciences and Engineering Research Council of Canada, or Social Sciences and Humanities Research Council.

This material is based in part upon work supported by the National Ecological Observatory Network (NEON), a program sponsored by the U.S. National Science Foundation (NSF) and operated under cooperative agreement by Battelle. Data used in this research were provided by the [PhenoCam Network](https://phenocam.nau.edu/webcam/), which has been supported by the National Science Foundation, the [Long-Term Agroecosystem Research (LTAR)](https://ltar.ars.usda.gov/) network which is supported by the United States Department of Agriculture (USDA), the U.S. Department of Energy, the U.S. Geological Survey, the Northeastern States Research Cooperative, and the USA National Phenology Network. We thank the PhenoCam Network collaborators, including site PIs and technicians, for publicly sharing the data that were used in this paper.

We would also like to thank Daniel Rubenstein for the guidance, and Mike Long and his team for the generous assistance during the field work. 

<!-- You may also want to credit the source of your data, i.e., if you went to a museum or nature preserve to collect it. -->

<!-- ## Glossary  -->

<!-- [optional] If relevant, include terms and calculations in this section that can help readers understand the dataset or dataset card. -->

<!-- ## More Information  -->

<!-- [optional] Any other relevant information that doesn't fit elsewhere. -->

## Dataset Card Authors 

Jianyang Gu

## Dataset Card Contact

Jianyang Gu (gu.1220[at]osu[dot]edu)
<!-- Could include who to contact with questions, but this is also what the "Discussions" tab is for. -->