WHAT EXACTLY DOES IT DO?

South Africa boasts remarkable biodiversity, but it is currently facing severe threats from human activities, making ecological conservation imperative. The identification of biodiversity hotspots in South Africa can significantly contribute to environmental preservation in the region. However, these hotspots have remained largely uncharted until now.

To comprehensively explore these unknown regions, we use high-spectral data that was previously unavailable and conduct spectral analysis with recent satellite imagery from the EMIT project. Through this analysis, we have identified 15 potential hotspot locations.

How does it work?

We conduct spectral analysis using data from the Earth Surface Mineral Dust Source Investigation (EMIT) project. Firstly, we utilized the addition and subtraction of energy in characteristic wavelength segments as indicators to detect specific disturbances, such as agricultural land (NDMI ((842nm-1610nm)/(842nm+1610nm)) > 0.2), clouds, sandy areas, water bodies (NDWI ((560nm-842nm)/(560nm+842nm)) > 0), buildings, and areas with insufficient vegetation (NDVI ((842nm-665nm)/(842nm+665nm)) < 0.35).

Subsequently, we analyze spectral diversity on the remaining areas. First of all, we assumed that areas with high diversity make up a very small proportion of the total grid points, and the points with the highest spectral diversity would exhibit the greatest biodiversity. This is because different plant characteristics result in different reflection spectra. Therefore, we calculated the Z-score for each band at each grid point, determining the differences in the same band among different grid points. Finally, we calculated the absolute value of average of the cumulative Z-scores for each grid point to obtain a distribution of grid point variability. The grid point with the highest Z-score represents the most dissimilar spectral signature from others, indicating the highest spectral diversity. According to our assumption, this represents the diversity in plant characteristics.

Finally, we deploy different layers to our website, providing researchers with a quick overview of the spatial distribution of various layers (NDMI, NDWI, NDVI, result1, result2) and potential hotspot locations.

See the github page for more detail:

https://github.com/huichu1005/map/tree/master

What benefits does it have?

We utilized the EMIT data to compute spectral diversity and devised a biodiversity hotspot selection method tailored for the South African region. Through this approach, we have identified potential hotspot locations. Compared to traditional fieldwork, the results of this study offer a more efficient and expedited avenue for biodiversity research in this area.

What do you hope to achieve?

Our goal is to identify areas with high biodiversity in the Cape region of South Africa and enhance the capacity for biodiversity conservation in the Cape region. We also aim to raise awareness among the public about the importance of biodiversity conservation.

Tools, Languages, Hardware

【LANGUAGES】

Python is our coding language for analysis.

HTML, CCS, Javascript is for website deployment.

We speak Mandarin.

【HARDWARE】

24GB RAM

【SOFTWARE】

Visual Studio Code for coding

Microsoft PowerPoint for presentation slides

Github for deployment

Bootstrap v5.3.2

jQuery v3.6.0

tailwind

The python library list one of our members used:

- xarray=2023.6.0

- folium=0.14.0

- geopandas=0.13.2

- matplotlib-base=3.7.1

- jupyter=1.0.0

- numpy=1.24.4

- hvplot=0.8.4

- branca=0.6.0

NASA python library: emit_tools

2023 NASA Space Apps Challenge

$quality_over_quantity