Phase 1: Aquatic reflectance validation ======= This phase focusses on the radiometric validation and app usability. Research groups working on aquatic reflectance data collection and analysis are invited to collaborate. This research phase is registered [here][0] Timeline: This project phase will be launched at the Ocean Optics conference 2024 in Las Palmas, followed by further outreach through professional networks. This phase is intended to conclude one year from launch. Hypotheses --- - Aquatic reflectance can be derived from measurements of water, sky and grey card radiance at 135 degrees from solar azimuth and 40 degrees from vertical, using the iSPEX 2 spectropolarimeter. - Accuracy will be within 25% (threshold), 15% (breakthrough), 10% (target) at the peak sensitivity of the red, green and blue range of smartphone cameras, compared to reference measurements taken in field conditions. - A signal to noise ratio > 100 (threshold), > 1000 (target) can be achieved, measured at representative wavebands. - Spectral aberrations resulting from polarisation filters can be effectively removed by demodulation algorithms and any subsequent non-linear fitting. Research questions --- Our main research questions at this stage are: - Optimal methods to correct for smile effect - Signal-to-noise ratio obtained in radiance measurements - Propagation of image calibration methods to new phones - Reproducibility in radiance (incl. grey card) and reflectance - Demodulation algorithms Practical functionality being tested at this stage: - Reproducible data collection (with basic training) - App stability - Data reporting (from reference sensors, including matching by location/time) - Spectral calibration (using CF lamps) Design Plan --- Observational Study: Data are collected from study subjects that are not randomly assigned to a treatment. This includes surveys, “natural experiments,” and regression discontinuity designs. Study design: At this stage of the project, we are launching a campaign for researchers in the water optics community to join the iSPEX 2 validation effort. Our aim is to publish a paper detailing the performance and use cases of iSPEX 2 in registering water colour following approximately one year of data collection efforts. Contributors (> 10 samples) would be co-authors of this paper, and involved in the research steps below. iSPEX 2 units will be provided (1 per group or location), compatible phones are expected to be provided by the research teams. Positioner plates will be 3d printed and should be fixed in place to prevent any light leakage (injection moulded designs, used in future, should not suffer light leakage). At this stage of the project, the iSPEX 2 App is still undergoing functional and design changes. The focus of data analytics is on quality control during processing steps leading to aquatic reflectance. Collaborators contributing substantially to data collection, quality control and analysis will be invited to co-author the resulting manuscript with the working title 'Validation and uncertainty estimation of smartphone-based aquatic reflectance using iSPEX 2' Data collection procedures --- Participants are now invited to inform project leads of their interest to participate. They will then plan how to obtain iSPEX 2 for a specific phone model, which will be mailed to them. Data collection is to be alongside and synchronous with reference measurements, capturing the same illumination conditions. Data collection will follow the 'Mobley protocol' at 135 degrees azimuth and 40 degrees oblique angle. Obstruction of the downwelling light field is to be prevented, including measurement of the grey card. Participants will keep a log of their observation windows and provide the nearest-in-time matching reference radiometric measurements (Lt, Ls, Ed or Lr) using a template, so that data can be automatically ingested by post-processing and data analysis scripts. Access to the project repository will be extended to participants to provide data updates and corrections. Sample size --- We anticipate working with around 20 research teams, providing at minimum 10 observation days each. There is no maximum size to the data set generated. The maximum number of participating groups is around 50, limited by the number of iSPEX 2 units presently available. The minimum number of observations is set to provide a representative sample for regression analysis. Variables --- We expect to collate a globally representative sample of inland, coastal, ocean water types. - Water, sky and grey card radiance using iSPEX 2 - Water, sky radiance and downwelling irradiance using reference sensors. - Reference sensors will be 'hyperspectral' providing sufficient resolution to simulate the response of iSPEX 2. Reference sensors will measure at the same viewing angles as considered optimal for iSPEX 2. Indices --- Samples will be grouped in observation pairs (nearest in time < 1 minute) and further by station / location, UTC timestamp accurate to 1s, and location (lat/lon) accurate to 6 decimals. Samples will further be annotated with iSPEX 2 serial number, and an observer identifier. Analysis Plan --- Python scripts will be developed to analyse the incoming data. Type-II regression models will form the main component of spectral validation, following essential quality controls, and mapping of hyperspectral radiance to equivalent spectral response functions (or, at least, wavelength coordinates). Reconstruction of reflectance using foward models may be used to further determine the integrity of the observations. Future research questions === - propagation of measurement uncertainty - algorithm development and tuning for key water quality properties - atmospheric optical thickness measurements Data sharing: - regular data dumps (analysis-ready) in the code repository - analysis scripts maintained in the code repository - results posted on the OSF project page. [0]: https://osf.io/m5qwd