Chlorophyll a and turbidity distributions: applicability of using a smartphone ‘‘app’’ across two contrasting bays

Turbidity from suspended particulate matter and chlorophyll a concentrations are first-level indicators of the health of coastal waters. Coastal regions, typically several kilometers in scale, require a sampling resolution sufficient to identify the extent of turbidity and chlorophyll a sources and sinks. An inexpensive smartphone “app” (HydroColor™), which uses color reflectance to measure turbidity and chlorophyll a concentration, was tested for both accuracy and practicality. Two separate and contrasting coastal regions were chosen, the proposed Silam Coastal Conservation Area, and Sepanggar Bay, serving Kota Kinabalu city (Sabah, Malaysia). It was found that results produced by HydroColor's built-in turbidity calibration agreed well with those from a laboratory nephelometer across the range encountered within the two regions (0 to 26.4 nephelometric turbidity units). Chlorophyll a concentrations, calculated using a function developed with a previous less-convenient underwater camera setup, showed a response closely proportional to chlorophyll a water extractions, but greater by 1.30 μg L⁻¹. Nevertheless, the differences and similarities between turbidity and chlorophyll a patterns between bays were compelling, especially as they were consistent with land use and river and coral reef proximity. The study identified several practical difficulties and sampling design constraints. The extent of the survey was limited to regions of the bay where the bottom could not be seen. Large tidal variations required three small, fast boats for rapid random site visits before onset of sea breeze to normalize for tide and wind variability. Reflectance measurements and accuracy were also hampered by screen glare, avoiding changing cloud cover, and rain staining the reflectance grey card, as well as the difficulty in avoiding interference from shadows from restricted open deck space. In contrast, the accuracy and precision of the regional calibration appeared to be dependent on obtaining a sufficient number and range of values over different turbidity and cloud conditions across the region and over different times.