2020 ESA Annual Meeting (August 3 - 6)

PS 20 Abstract - Reckless abandonment? assessing how agricultural legacies influence ecosystem recovery

Damien Farrant1, Ashley E. Larsen1 and Dar A. Roberts2, (1)Bren School of Environmental Science and Management, University of California Santa Barbara, Santa Barbara, CA, (2)Department of Geography, University of California Santa Barbara, Santa Barbara, CA
Background/Question/Methods

Millions of hectares of agricultural land have been abandoned globally since the mid-20th century. Abandoned land may support vegetation similar to its pre-agricultural state, offering biodiverse habitat and a wealth of ecosystem services. Alternatively, abandoned land may foster degraded ecosystems dominated by invasive species. Despite widespread and increasing abandonment, we still lack an understanding of when and where biodiversity may benefit from abandonment. This study examines vegetation succession patterns on abandoned agricultural fields in Hawai‘i where cultivated area decreased by 50% (80,000 ha) from 1980 to 2015. The steep environmental gradients, a plethora of endemic species, and legacy of abandonment following plantation closures position Hawai‘i as an ideal site to study the effects of various environmental contexts on succession trajectories. Changes in land cover were detected on fourteen Landsat satellite images between 1989 and 2018. Hierarchical decision tree classifiers were developed in R based on the supervised selection of reference targets in three satellite images. The decision trees were used to classify land cover in all of the satellite images in ENVI. Here we address two questions: to what extent do abandoned lands revert to native ecosystems?; And what environmental contexts give rise to native versus invasive species succession trajectories?

Results/Conclusions

We find that ecosystems dominated by invasive species are prevalent on abandoned land, occupying 50% of the total abandoned area with patches present on the six largest Hawaiian Islands. Native ecosystems have succeeded to a lesser extent, covering 28% of the abandoned fields. Concentrated clusters of native ecosystems occur on Molokai and Lana‘i, while large and small patches are dispersed around Hawai‘i and Maui, respectively. Preliminary findings indicate that native ecosystems favor higher elevations and former pasture land, where more than 64% of the native ecosystems that have reestablished on abandoned land are found. Ongoing analysis of these succession patterns in tandem with key landscape variables (e.g. soil type and residual pesticide compounds) in a regression discontinuity analysis framework could assist in characterizing more nuanced threshold conditions that determine succession trajectories. Insight derived from this study in Hawai‘i can guide a broader and more global effort to manage land more sustainably with an improved understanding of the conservation opportunities or threats that abandonment may present to ecosystem biodiversity.