Ocean Acidification

For the first time, the Autosub Long Range (ALR) completed a fully autonomous, long-distance (2,000 km) scientific mission, delivering new insight into coastal carbonate dynamics and air-sea CO2${\text{CO}}_2$ fluxes. Equipped with a suite of oceanographic sensors, including a Lab-on-Chip (LOC) pH sensor and a Sea-Bird SeaFET pH sensor, the mission generated nearly 50,000 high-resolution pH measurements, providing one of the most detailed continuous coastal carbonate data sets collected to date in the region. We evaluated the adjustment of the SeaFET reference potential (k₀), testing both the co-deployed LOC sensor and neural network estimates as reference pH. Before correction, the LOC and SeaFET sensors showed close agreement (Δ$\mathrm{\Delta }$pHT${\text{pH}}_{\mathrm{T}}$ = 0.013 ±$\pm$ 0.009), which improved to Δ$\mathrm{\Delta }$pHT${\text{pH}}_{\mathrm{T}}$ = 0.00004 ±$\pm$ 0.007 after LOC-based k₀ adjustment. Both sensors diverged from model estimates, indicating reduced ability of models to resolve fine-scale coastal variability and reinforcing the need for direct in situ observations. Total alkalinity (TA) was derived from salinity-based relationships and model predictions, and paired with pH (SeaFET, LOC, and modeled) to estimate the partial pressure of CO2${\text{CO}}_2$ (pCO2${\text{CO}}_2$), which ranged 263–598 ±$\pm$ 27 μatm. Resulting air-sea CO2${\text{CO}}_2$ fluxes ranged −17.0 to 7.1 ±$\pm$ 1.09 mmol m−2 d−1${\mathrm{m}}^{-2}{\mathrm{d}}^{-1}$, with the Celtic Margin acting as a net CO2${\text{CO}}_2$ sink in May–June of 2022. pCO2${\text{CO}}_2$ and CO2${\text{CO}}_2$ flux proved sensitive to subtle pH differences, but less so to TA estimates. Our findings demonstrate the critical role of high-resolution autonomous observations in quantifying coastal carbonate dynamics and CO2${\text{CO}}_2$ fluxes, capturing processes and variability that are largely unresolved by ship-based surveys or global models.

Plain Language Summary

For the first time, the Autosub Long Range (ALR), a robotic underwater vehicle, completed a long-distance (2,000 km) scientific mission without ship support. During the mission, the ALR collected ocean data using sensors that measured water temperature, oxygen, and seawater acidity (pH). The mission generated nearly 50,000 high-resolution pH measurements, providing one of the most detailed continuous coastal carbonate data sets collected in this region. We evaluated how well two onboard pH sensors agreed with each other and with model-based estimates, and tested an applied correction method. Before and after correction, the sensors closely agreed, showing that autonomous platforms can gather high-quality chemical measurements that resolve coastal marine carbon dynamics better than model estimates alone. We used the pH data, along with estimates of TA, to calculate carbon dioxide (CO2${\text{CO}}_2$) exchange between the ocean and atmosphere, finding that the region operates primarily as a carbon sink in May–June. Our calculations were strongly influenced by pH input but much less by alkalinity, highlighting the importance of accurate pH observations. This study shows that autonomous sensors can improve monitoring of carbon cycling in dynamic, understudied coastal waters that play a key role in the global carbon system.