Author: Michela Perrone
Nearly three decades ago, the Spectral Variation Hypothesis (SVH) proposed a tremendously simple idea: the spectral diversity of vegetated landscapes (essentially, the spatial variation in reflectance detected by remote sensing) could serve as a proxy for plant species richness. The promise of biodiversity mapping from space quickly gained traction, especially with multispectral satellite data offering an efficient alternative to traditional field-based monitoring. But as remote sensing technologies advanced, so did the scrutiny of the SVH. What once seemed like a straightforward link between spectral heterogeneity and biodiversity turned out to be far more complex.
Conceptual scheme of the Spectral Variation Hypothesis
Indeed, in recent years, several studies (including our own work) have explored how different approaches to spectral diversity affect the link with biodiversity at various ecological scales, revealing that spectral differences at the leaf level do not necessarily scale up to whole ecosystems. The findings of these studies reinforce a growing consensus: context matters. The effectiveness of spectral diversity has been found to depend on factors such as image characteristics, phenology, vegetation complexity and cover, indices used, habitat types, and sampling design. Such factors, thus, introduce considerable noise into the relationship between spectral diversity and biodiversity.
This disconnect raises a critical question: is it time to lay the SVH to rest? While some researchers argue for abandoning the hypothesis altogether, others suggest that SVH still has value.
At the recent BioSpace25 conference hosted by the European Space Agency (ESA) at its Centre for Earth Observation (ESRIN) in Frascati, Italy, Dr Christian Rossi, from the University of Zürich, put forward his view on the topic. In his perspective, even as the original SVH framework crumbles, certain aspects of spectral diversity continue to provide valuable insight and, while moving away from the initial simplicity of the SVH, there is space for leveraging more sophisticated methods and calculations of spectral diversity on large scales. Such refined approaches acknowledge the limitations of a one-size-fits-all model and instead incorporate additional parameters to enhance biodiversity estimations.
Some promising directions include:
- Sensor Fusion: Combining optical data with LiDAR or radar would allow for better characterization of vegetation structure, overcoming some of the limitations of spectral-only approaches.
- Temporal Analysis: Instead of relying on single-date spectral snapshots, monitoring spectral variation over time could capture phenological changes that relate to biodiversity.
- Sub-pixel Classification: Advanced classification techniques might help disentangle mixed spectral signals, improving single spectra discrimination.
- Uncertainty Quantification: Integrating error estimates into spectral models could provide a more realistic picture of biodiversity predictions
Thus, rather than discarding spectral diversity altogether, the key would be to move beyond simplistic assumptions and toward multi-faceted approaches. The SVH may no longer be the silver bullet it once seemed, but its ‘zombies’ (i.e., refined and integrated spectral techniques) are believed to help shape the future of biodiversity monitoring. With advances in remote sensing, machine learning, and ecological modelling, the next generation of biodiversity assessment tools will likely embrace a hybrid strategy, combining spectral data with complementary environmental and functional variables. This evolution in thinking marks a significant shift, from searching for a universal spectral-biodiversity link to tailoring approaches that consider scale, context, and ecological complexity.
The lesson from the SVH’s journey? Remote sensing remains a powerful tool for biodiversity science. By embracing complexity, we can ensure that the ‘zombies’ of the SVH continue to contribute to ecological research and conservation efforts worldwide.
This blog post is based on findings from our recent research and discussions on the future of spectral diversity in biodiversity monitoring. If you’re interested in learning more or collaborating, feel free to reach out!
For a thorough overview of the SVH and its challenges: Torresani, M., Rossi, C., Perrone, M., Hauser, L. T., Féret, J. B., Moudrý, V., … & Rocchini, D. (2024). Reviewing the Spectral Variation Hypothesis: Twenty years in the tumultuous sea of biodiversity estimation by remote sensing. Ecological Informatics, 102702. https://doi.org/10.1016/j.ecoinf.2024.102702