From the smallest microbes wriggling in the soil to the largest mammals roaming the grasslands, biodiversity is vital to maintaining the stability of ecosystems. It underpins global food security and planetary health. As the urgency to protect biodiversity grows, organizations and governments are mobilizing by setting ambitious goals, crafting policies, and developing market instruments like crediting schemes. While these efforts are encouraging, they also highlight the need to rapidly deploy monitoring protocols to understand where to implement projects and to measure outcomes effectively.
But what exactly do we mean by “biodiversity monitoring”? This is like someone saying, "I'm interested in cooking food." Great! Say more. Are you making a full meal or a single dish? Do you have a specific cuisine or ingredient in mind? You get the idea. In the same way, interest in biodiversity monitoring requires specificity to be actionable. To effectively protect and monitor biological diversity, we must understand what we’re measuring, why these metrics matter, and the trade-offs of different monitoring regimes.
Biodiversity is complex and monitoring it is multi-faceted. Here are some examples that highlight the breadth of what biodiversity can refer to:
This list is not exhaustive, but it illustrates a few dimensions through which we can approach biodiversity monitoring. No single dataset captures the full picture, and evaluating a single site through multiple lenses may yield very different insights.
For example, an intact desert with native species will likely show lower species richness simply due to the fact that the ecosystem can support fewer plants and animals. Or a remote island might have low intactness but still serve as crucial habitat for an endangered lizard that lives nowhere else. These examples highlight the need to define clear monitoring goals: Is the priority to increase species richness, protect areas where the most threatened or range-restricted critters reside, maintain diversity within a single taxonomic group, or something else? Each metric offers unique insights that inform different conservation or management strategies.
The ability to remotely monitor biodiversity is crucial as it allows for cost-effective, scalable insights across vast and hard-to-access areas. And while remote sensing and machine learning technologies are indeed unlocking new possibilities for biodiversity modeling at large scales, they also come with limitations. Fundamentally, remotely sensed data cannot directly observe species presence or absence, with rare exceptions such as megafauna in open areas or emerging research in hyperspectral data. You simply can’t count the number of endangered geckos or identify understory fern species from satellite or aerial imagery alone. Instead, researchers must infer from field samples and habitat characteristics. Understanding how these datasets are created is key to determining their appropriate uses and the level of confidence or uncertainty to place in their outputs.
Because we cannot directly observe biodiversity remotely, we must rely on proxies—indirect indicators that suggest ecological conditions. These proxies can range from land cover classifications to vegetation health metrics, used to infer habitat suitability or potential species presence. While essential, these proxies are built on assumptions and generalizations that may not always hold true across different environments. For example, a map of forest cover might be used to indicate suitable habitat for a forest-dwelling lemur, but that doesn’t guarantee that you’ll find a lemur there. Proxies can offer useful insights but are inherently imperfect stand-ins for actual species-level observations and without careful validation, they can lead to misleading conclusions.
Ground-based sample data collection is a key input for biodiversity models, but this process is extremely expensive, time-consuming, and difficult to scale. Since it's impossible to sample every location at multiple points in time, we are forced to extrapolate from incomplete data and make assumptions about unsampled areas, introducing uncertainty. Additionally, this data is often historical—species observed in a given area years ago may no longer be present due to changes in habitat or climate. These factors place an upper limit on the accuracy and reliability of models, meaning they suggest where a species may reside but cannot guarantee its presence today.
Biodiversity datasets often rely on simplified categories that can obscure the full picture. For example, what defines an "intact" versus "fragmented" habitat, or a "native" versus "non-native" forest? These rigid categorizations can be misleading, as they fail to capture the nuanced, complex realities of natural systems. Ecosystems do not conform neatly to human-made boundaries, and our attempts to classify them into discrete categories can limit our ability to fully express and understand ecological diversity. The oversimplification inherent in these datasets must be acknowledged and carefully considered when interpreting biodiversity data.
Remote sensing data can be immensely powerful, but it is not a one-size-fits-all solution. Its value lies in its scalability and cost-effectiveness, and remote data can be a valuable starting point to determine where on-the-ground observations are needed for further validation. But remote data is a tool that can either empower or mislead, depending on how it is applied – both in the context of biodiversity monitoring and beyond. Users must tread carefully and be discerning about the limitations of a dataset and its suitability for a specific purpose. Careful vetting of proxies, thoughtful application of the data, and an understanding of its inherent uncertainties are essential to avoid drawing inaccurate conclusions when measuring the success of biodiversity initiatives.
To leverage biodiversity data effectively, it’s essential to choose the right lens (pun intended 😉). Here’s a sample of datasets available in the Lens Library and what they reveal:
Lens provides access to a wide range of biodiversity datasets, along with methodology documentation from the providers. By understanding the limitations of each dataset, comparing them, and viewing them alongside high-resolution imagery, users can gain insights that no single dataset can offer. With a range of data and imagery sources, Lens empowers users to monitor specific areas of interest more effectively. These combined perspectives form a crucial foundation for planning and prioritizing on-the-ground data collection or management actions, helping to uncover critical nuances and support more informed decision-making.
As we strive to meet ambitious climate and conservation goals, it’s important to define which aspects of biodiversity we’re measuring and why. Just as cooking a meal requires planning and combining the right ingredients, effective biodiversity monitoring depends on selecting the appropriate datasets and understanding their strengths and limitations. Remote sensing and other monitoring tools provide powerful insights, but no single dataset offers a complete picture. By combining multiple perspectives on biodiversity, we can better identify projects with the highest chances of success and strategically deploy resources in the future. This careful scrutiny deepens our understanding of the intricate, beautiful planet we call home and ensures that biodiversity efforts are as impactful as possible.
Log into Lens today to access our library of powerful biodiversity datasets, or reach out to us to learn more about how Lens can support your biodiversity monitoring.
.png)
%20copy.png)
