Simulation of lidar methods for ocean temperature sensing

The properties of the ocean are of wide interest in numerous fields. Measurements of oceanic temperature are particularly important due to their impact on global climate. However, in order to measure water temperature below the ocean surface, only in situ methods are currently available, with corresponding limits on sampling frequency. Light Detection And Ranging (‘lidar’) methods provide a potential solution. Oceanographic lidar systems use short pulses of light to interrogate the properties of the water column. Prior systems have used elastically-scattered light for applications including the measurement of turbidity, chlorophyll and fisheries, but it is also possible to use light which has undergone wavelength change via the inelastic process of Raman scattering. By measuring the polarisation and spectrum of Raman-scattered light, subsurface water temperature can be inferred from lidar returns. However, the best approach for a Raman lidar system is still unclear.

This work uses simulations to compare and assess the performance of several different approaches to temperature measurement with Raman lidar. This was accomplished by modifying an existing unpolarised Monte Carlo simulation to simulate polarised scattering and detection.

Two temperature measurement methods were studied. Methods involving the measurement of depolarisation were found to perform comparatively poorly: for good results, they required very small fields of view and clearest oceanic waters, and even then encountered problems due to multiple scattering. By contrast, methods using the ratio of Raman spectral bands performed much better, and allowed the development of a simple inversion algorithm to recover temperature from lidar returns. This will inform the design of future Raman lidar systems for temperature measurement.