Using Google Earth To Map The Distribution Of Mangroves In The Galapagos

Mangroves only occupy a tiny bit of the coastlines of the world, yet they are one of the most productive ecosystems on Earth. Mangrove forests provide important ecosystem services. They protect the coasts from erosion and catastrophes, like tsunamis. Mangroves’ intricate aerial roots and trunks are the nurseries of commercially important fish and lobster, as well as ecologically important fish, like sharks. They also help combat climate change by sequestering substantial quantities of carbon in the subsoil and their living structures. They are among the most carbon-rich tropical forests in the world.

Mangroves are threatened globally and, if current deforestation rates persist, they are expected to functionally disappear in 100 years. In Ecuador, mangrove deforestation rates have risen to as high as 40% in the last 40 years due to mangrove clearing for shrimp aquaculture. Thankfully, the Galapagos Archipelago is a World Heritage site with strict protection rules that kept the shrimp aquaculture away.

Are there mangroves in the Galapagos?

The Galapagos archipelago is not particularly known for its mangrove forests, yet there are mangroves in Darwin’s evolution’s theory backyard. The distribution of mangrove forests is considered critical baseline information, since knowing where there are mangroves is also the foundation of many other studies. Ecosystem services provision and evaluation, fisheries, shark biology, etc., as well as governance and spatial planning, need knowledge of the spatial extent of this important ecosystem. Several studies attempted to uncover the incognita of the distribution of mangrove forests in the Galapagos since 1986. But prior to this study, there was only an approximate knowledge of the distribution of mangroves in the Galapagos.

We used readily available Google Earth very high-resolution imagery to map mangrove distribution in the Galapagos. We digitized mangroves directly in Google Earth software and used a set of image interpretation attributes to distinguish mangrove from other features. We determined that there are 3657 ha of mangroves covering 35% of the coastline of the archipelago. To evaluate the classification accuracy, we used ground-truth data form 208 sites and obtained an overall classification accuracy of 99.1% and a Kappa of 0.97, surpassing all previous studies.

The interesting bit is that, once digitized, we can perform all sorts of spatial analysis. We propose a set of measures to evaluate mangrove patch complexity that complements the comparison between studies and can also be used to assess the digitization process.

How it compares to other methods, though

We wanted to be sure that this was the best method to map the mangroves of the Galapagos. So we compared our method with other, more traditional ones in the field of vegetation cover classification. To do so, we needed to compare apples to apples, so we downloaded Google Earth imagery and applied semi-automatic classification algorithms, namely the Maximum Likelihood Classification (MLC) algorithm, and object-based image analysis, coupled with MLC. We tested the accuracy of each of the methods using ground-truth data scattered across the archipelago. In either case, on-screen digitization proved to be superior to semi-automatic classification algorithms, probably because these work well when there is enough spectral information. However, Google Earth imagery only has information in the visible spectrum (Red, Green, and Blue wavelengths).

A cost-effective solution

We ensured the results obtained using Google Earth very high-resolution imagery were the most accurate. But, what about the costs of producing the mangrove cover compared to other methods such as using drones?

We showed that our method is between 400 to 35,000% cheaper than other solutions in the Galapagos, which has a huge area to be mapped (around 4000 km2).

Google Earth’s very high-resolution imagery is indeed a very interesting solution for anyone attempting to map mangrove forests or any vegetation cover, granted that it can be easily distinguishable from other vegetation. We acknowledge this method is particularly useful for semi-arid to arid tropical regions of the world, such as Baja California in Mexico, the Red Sea, the mangroves of NE of Brazil, New Caledonia, etc., which are also characterized by being composed of small patches. In the Galapagos, we found that 80% of mangroves have less than 0.25 ha. This poses a difficulty in their detection with remote sensing techniques because they need to have a high enough resolution to detect them.

However, the most used, freely available satellite imagery, like Landsat, have a resolution that leaves these small patches virtually undetected. Hence the need for really high-resolution imagery. The possibility to use Google Earth high-resolution imagery for this matter is very good news to Marine Protected Areas (MPA) in poor and middle-income countries, or simply in MPAs with little funding capacities to get expensive imagery sources. Protected areas and governments can monitor and study mangrove cover and how it is changing over time using an affordable source of information.

Mangroves are thriving globally, but is the protection of the Galapagos World Heritage working?

Galapagos’ mangroves are one of the very few examples of mangroves in the world that are virtually free of human impacts, at least from local impacts. This means that there is no mangrove deforestation but also no mangrove reforestation. We showed that in a ten-year time span, the overall growth of mangroves has been of 24%. Other studies found similar mangrove expansion rates, however, these are the result of the balance between the extraction of mangrove forests and the afforestation by restauration programmes. Another interesting finding is that this ten-year growth is mainly the result of the expansion of existing mangrove patches and, to a lesser extent, the generation of new patches that did not exist before. We also found that there is very little disappearance of mangrove patches, the losses are mainly due to the contraction of patches. Now that we determined these patterns, we need further studies to understand the processes behind the patterns.

By analyzing the vegetation cover of the coastline and relating it to the geological age of each island, we could estimate the relationship between them. Mangrove cover and mangrove growth are inversely proportional to the geological age of the islands, which apparently supports the idea of mangroves as pioneering coastal vegetation in the archipelago. However, there are many other factors, in addition to island age, that might influence mangrove distribution such as nutrients, currents or wave exposure.