Light data was collected from a set of underwater light sensors, a combination of underwater and above-water sensors, or vertical profilers. The exact setup of the sensors varied across lakes. Sensor models used include LI-COR, Onset, Odyssey, Apogee, and Sea-Bird light sensors. Onset sensors measured light in lux, while all other models measured photosynthetically active radiation (PAR) in micromoles per square meter per second. Raw light values were converted to light attenuation values according to the methods below, which are taken directly from the accompanying manuscript:
To convert light measurements from paired sensors into light attenuation, we used the following equation (Kirk, 1994):
Kd=1/(z2-z1 )ln(Ed(z1))/(Ed(z2 ))
where Ed(z) is the measurement of downward irradiance at depth z and z2>z1. In the cases where more than 2 underwater light sensors or a vertical profiler was used (16 lakes), Kd was calculated as the slope of the fitted regression line of ln(Ed) vs depth. Light measurements were obtained as PAR (umol/m2/s), Lux (lumens/m2), or wavelength-specific measurements (uW/cm2/nm). Although Lux and PAR both span 400 nm to 700 nm in the electromagnetic spectrum, they are not equivalent spectra. While Lux can be converted to PAR by calibration using in situ measurements (Long et al., 2012), only 4 of the 17 lakes using Lux measurements had this conversion available. Therefore, Lux measurements were not converted to PAR for the purpose of Kd calculation. A detailed treatment of potential limitations and assumptions on sensor design and deployment is included in the supplemental section of this paper. For comparison, wavelength-specific measurements were converted to PAR as the sum of all light measurements between 400 nm and 700 nm before calculating Kd.
To ensure calculations were minimally influenced by the effects of sun angle and low-light conditions, we only used light measurements taken within 3 hours of local solar noon. While sun angle will impact light attenuation based on season and latitude (Kirk, 1994), this was not found to affect daily rates of change and the effects of seasonal sun angle were not considered in this analysis, except for the consideration of albedo when necessary. For lakes with an above-water sensor, we adjusted surface irradiance to irradiance just below the surface according to Paulson and Pegau (2001):
Ed^(0-)=(1-a)Ed
where Ed^(0-) is irradiance just below the surface, a is albedo, and Ed is downwelling surface irradiance. We estimated albedo based on the solar zenith angle (zen) according to Briegleb et al., 1984:
a = 1/100(2.6/(1.1cos(zen)^1.7+0.065) +15[cos(zen )-0.1][cos(zen )-0.5][cos(zen)-1])
We calculated all values of solar zenith angle based on local time and latitude using the R package suncalc (v. 0.5.0). Interferences such as wave focusing, buoy shadows, and wiper placement can affect light attenuation calculations. To reduce their effect on daily light attenuation estimates, after calculating Kd for each time step, we removed data points where the ratio of irradiance at the lower light sensor to that at the upper light sensor was outside the first and third quartile of the distribution for each day’s observations. Data points where more irradiance was measured at the lower sensor than the upper sensor were then removed, obvious erratic measurements were manually removed (i.e., measurements where sensors were not underwater or were affected by fouling), and the remaining measurements were used to calculate light attenuation as a daily average. Days were removed from further analyses if more than 75% of the measurements within the daily six-hour window were removed by this process (n = 443 days), and the remaining days were used to generate a time series.
References:
Briegleb, B. P., Minnis, P., Ramanathan, V., & Harrison, E. (1986). Comparison of regional clear-sky albedos inferred from satellite observations and model computations. In Journal of Climate & Applied Meteorology (Vol. 25, Issue 2, pp. 214–226). https://doi.org/10.1175/1520-0450(1986)025<0214:CORCSA>2.0.CO;2
Kirk, J.T.O. (1994). Light and photosynthesis in aquatic ecosystems, third edition. Cambridge University Press. https://doi.org/10.1017/CBO9781139168212
Long, M. H., Rheuban, J. E., Berg, P., & Zieman, J. C. (2012). A comparison and correction of light intensity loggers to photosynthetically active radiation sensors. Limnology and Oceanography: Methods, 10(JUNE), 416–424. https://doi.org/10.4319/lom.2012.10.416
Paulson, C. A., & Pegau, W. S. (2001). Penetrating shortwave radiation. Encyclopedia of Ocean Sciences, 4, 379–384. https://doi.org/10.1016/B978-012374473-9.00154-5