Spatial Averaging of Oceanic Rainfall Variability Using Underwater Sound. Ionian Sea Rainfall Experiment 2004: Acoustic Component

An experiment to evaluate the inherent spatial averaging of the underwater acoustic signal from rainfall was conducted in the winter of 2004 in the Ionian Sea southeast of the coast of Greece. A mooring with four passive aquatic listeners (PALs) at 60, 200, 100, and 2000 m was deployed at 36.85 deg. N, 21.52 deg. E, 17 km west of a dual-polarization X-band coastal radar (XPOL) at Methoni, Greece. A dense rain gauge network was set up in Finikounda, 10 km east from Methoni to calibrate the radar. Eight rain events were recorded during the deployment; six of these events were recorded by both the PALs and the XPOL radar. The acoustic signal is similar at all depths and rainfall was detected at all depths, although the deeper PALs suffered an unexpected sensitivity loss and consequently did not trigger into high sampling mode as often as the shallower PALs. The total accumulation reported is lower for the deeper PALs. The acoustic signal is classified into wind, rain, shipping, and whale categories. The shipping signal increased throughout the deployment. A signal from whales is present roughly 2% of the time, most often at 200 m, and is consistent with the clicking of deep-diving beaked whales, although there was no visual confirmation of whale presence. Acoustic co-detection of rainfall with the radar demonstrates the need to classify the rainfall signal in the presence of other underwater noises. Once detection is made, the correlation between acoustic and radar rainfall rates is high. Spatial averaging of the radar rainfall rates in concentric circles over the mooring shows highest correlations with increasing acoustic recording depth, verifying the larger inherent spatial averaging of the rainfall signal with recording depth. For the PAL at 2000 m, the maximum correlation was at 3-4 km, suggesting a listening area for the acoustic rainfall measurement of roughly circular area = 30-50 km2, in contrast to less than 3 km2 for the acoustic measurement at 60 m depth.