Multi-decadal variability in sandy beach area and the role of climate forcing

Abstract

Sandy beaches comprise 31% of the ice free coasts of the world, providing a variety of ecosystem services essential to support human well-being, particularly protection against erosion and extreme climatic events. These highly dynamic environments are controlled principally by wave energy, tides and grain size, generating complex morphodynamic patterns that translate into strong area variation. Estimates of beach area and its variability are prime inputs for coastal management, a challenge that requires data representing a wide spatiotemporal scale. In this paper, a Landsat-based semi-automated methodology was developed to reconstruct the area of 21 sandy beaches of the Montevideo (Uruguay) coast from 1984 to 2016. This long-term information was also used to discern erosion-accretion cycles by means of wavelet analysis, and to explore the role of climatic forcing on these cycles through linear mixed models. A random forest classification algorithm was applied to Landsat images in order to estimate beach area. Long-term trends described a 27-year cycle with well-delimited quasi-decadal erosion and accretion phases related to climatic configurations. The beach area was negatively affected by an increase in sea level and climatic conditions of the previous year, being positively correlated with sea surface temperature anomalies and offshore winds (which favored accretion) and negatively correlated with onshore winds and intense El Niño Southern Oscillation (ENSO) events (favoring erosion). Our findings, together with the predicted increase in the occurrence and intensity of storms and extreme ENSO events, constitute a worrying scenario, as erosion in this populated coastal zone could have negative social, economic and ecological repercussions. The methodology developed here was useful to detect long-term changes in beach area and is entirely based on open-access information. Therefore, it is potentially applicable at any location on the planet. This approach is also useful to counteract the scarcity of long-term information that has precluded robust assessments of climate change effects on coastal zones.