In recent years, blue carbon ecosystems, in particular mangroves, have received unprecedented attention as a means to address climate change and biodiversity loss. While less in the spotlight, seagrass has also garnered interest, owing to its carbon sequestration capabilities and critical importance to ecosystems and local communities. AI and remote sensing is now being applied to these ecosystems to address their specific challenges.
Seagrasses in their entirety have greater carbon mitigation potential than mangroves due to their global abundance. Seagrass meadows potentially contain 4,350-8,550 Tg of organic carbon in bed sediment and biomass. If degraded or destroyed, they would be a significant carbon source. Further, they provide food and habitat for biodiversity including dugongs, manatees and sea turtles and important ecosystem services, such as food security, protection from coastal erosion and buffering against floods.
Seagrasses are a diet staple for many threatened species, including dugongs, manatees and sea turtles. Photo credit: Julien Willem (2008)
However, seagrass conservation and restoration projects face several challenges including the fact that they require significant financial and logistical resources. Effective protection of seagrasses involves protecting systems connected to the seagrass meadows such as preventing or reducing pollution in waterways that drain into seagrass habitats. Planting seeds or transplanting seedlings can also be logistically demanding and they are susceptible to loss and degradation. Situated in the intertidal zone, seagrasses are affected by both marine and terrestrial drivers of change. The permanence of their carbon storage is also difficult to ensure in the face of storms, heatwaves and invasive species. Additionally, in the context of carbon offsets, demonstrating additionality for seagrass projects can be challenging. The paucity of data on historical seagrass extent and the inherently patchy and dynamic nature of seagrass landscapes make it difficult to establish a robust baseline. Proving the effectiveness of the project to maintain or increase seagrass extent is another challenge.
Applying remote sensing and AI to seagrass can address a number of these challenges. One key area would be in improving the cost-effectiveness and efficiency of project planning, implementation and monitoring. Traditional methods of mapping seagrass habitats, such as visual surveys by scuba divers or towed cameras, are time-consuming, expensive, and often capture a limited scope of data. Conversely, remote sensing can assess large areas of seagrass efficiently. Machine learning algorithms can be trained to identify the spectral signatures, surface characteristics and other attributes of seagrass to distinguish them from other types of marine vegetation and seafloor substrates. Applying AI in tandem with remote sensing eliminates the need for manual interpretation, streamlining the processing of spatial data. This allows for efficient identification of seagrass habitats, health and productivity assessment, and insights into spatiotemporal change dynamics to inform management decisions.
Critically, this efficiency does not come at the expense of accuracy. Remote sensing and AI are especially useful for seagrass habitat mapping and density assessments, with some studies reporting accuracy scores of up to 95%. While accuracy can vary widely across geographies, the range in reported scores owes to differing site conditions and inherent limitations of a submerged system, rather than methodological capabilities. Using in-situ measurements to train and validate machine learning algorithms can improve the accuracy of the model, while the application of AI allows for analysis to be performed at scale (check out our article on field measurements and remote sensing).
The discourse around seagrass is as complex as the ecosystem of interest. Despite this, the challenges surrounding seagrass projects are patently overshadowed by the significance of seagrass to coastal communities and biodiversity. Cognizant of this, there has been growing interest in developing seagrass projects as well as the value of remote sensing and AI. The application of remote sensing and AI to seagrass landscapes is not without limitations and does not resolve all the challenges surrounding seagrass projects. However, they have proven and will continue to advance our understanding and efforts to protect and restore this valuable ecosystem.
Comments