Fractional coverage rather than green chromatic coordinate is a robust indicator to track grassland phenology using smartphone photography

Abstract

Cost-effective and accurate methods are critical in monitoring grassland phenology. Smartphones have great potential for tracking phenological changes owing to their easy operation and flexible sampling, especially for experimental treatments on grassland phenology at local scale. However, it remains an open question for an accurate and robust measure in tracking plant phenology under various light conditions when using smartphones. To fill this gap, we compared green chromatic coordinate (GCC) and fractional coverage (FC) produced by three widely-used smartphone types across varying light conditions. We also explored the effects of experimental warming and N addition on grassland spring phenology using smartphone photography. We found that light conditions affected GCC values derived from various smartphones. The percentile method did not reduce the influence of various light conditions on GCC calculation. When using one type of smartphone, other techniques (e.g., shading strong sunlight) were required to minimize the effect of light on mean GCC during photography, but they did not work for FC. Based on this method, three smartphone types generated similar phenological metrics (i.e., start, peak and end of growing season) detected from GCC or FC. GCC and FC produced synchronous seasonal trajectories and similar phenological metrics throughout the growing season. We also found that experimental warming advanced spring phenology in the relatively wet year of 2018, but delayed it in the dry year of 2019. In contrast, N addition had little influence on spring phenology. Our results revealed that overlooking the influence of ambient light and smartphone types on GCC might lead to uncertainties in phenological measurement. Robust FC metrics extracted from smartphone photography detected grassland phenology differences in response to warming and N deposition. These findings provide evidence that FC is a robust indicator to accurately monitor grassland phenology using smartphone photography. Their utility resonates with those monitoring many experimental plots or sites with various light conditions.