|Published in||Geophysical Research Letters, 47, e2020GL089093|
Meinen, C.S.,  Perez, R.C., S. Dong, Piola, A.R., & Campos, E.J.D.
1NOAA Atlantic Oceanographic and Meteorological Laboratory, Miami, FL, USA, 2Servicio de Hidrografía Naval, Buenos Aires, Argentina, 3Universidad de Buenos Aires, Buenos Aires, Argentina, 4Instituto Franco‐Argentino sobre Estudios de Clima y sus Impactos, Consejo Nacional de Investigaciones Científicas y Técnicas, Buenos Aires, Argentina, 5Oceanographic Institute, University of São Paulo, São Paulo, Brazil, 6Department of Biology, Chemistry and Environmental Sciences, School of Arts and Sciences, American University of Sharjah, Sharjah, United Arab Emirates
Christopher S. Meinen1 , Renellys C. Perez1 , Shenfu Dong1 , Alberto R. Piola2,3,4 , and Edmo Campos5,6
• Decade‐long hourly records of ocean bottom temperature demonstrate a
surprising amount of variability at a range of time scales
• Two deep/abyssal Argentine Basin sites (>4,500 m) exhibit significant warming trends of 0.02°C ± 0.01°C per decade over the period 2009–2019
• Sampling at least once a year is necessary to capture deep/abyssal temperature trends with at least ±50% accuracy (95% confidence limits)
Consecutive multi‐year records of hourly ocean bottom temperature measurements are merged to produce new decade‐long time series at four depths ranging from 1360 m to 4757 m within the northwest Argentine Basin at 34.5°S. Energetic temperature variations are found at a wide range of time scales. All sites exhibit fairly‐linear warming trends of approximately 0.02‐0.04°C per decade over the period 2009‐2019, although the trends are only statistically different from zero at the two deepest sites at depths of ~4500‐4800 m. Near‐bottom temperatures from independent conductivity‐temperature‐depth profiles collected at these same locations every 6‐24 months over the same decade show roughly consistent trends. Based on the distribution of spectral energies at the deepest sites and a Monte Carlo‐style analysis, sampling at least once per year is necessary to capture the significant warming trends over this decade to within 50% error bars at a 95% confidence limit.