Abstract: Determining soil water content (SWC) is fundamental for soil science, ecology and hydrology. Many methods are put forward to measure SWC, such as drying soil samples, neutron probes, time domain reflectrometry (TDR) and remote sensing. Sampling and drying soil is time-consuming. A neutron probe cannot determine SWC of surface soil accurately because neutrons escape when they are emitted near soil surface and TDR is, to some extent, influenced by soil salinity and temperature. Remote sensing can obtain SWC over a large area across a range of temporal and spatial scales. Complicated terrain and atmospheric conditions often make remote sensing data unreliable. Determining SWC from surface gray level (GL) measurements in the visible spectrum may have advantages over other remote sensing techniques, because surface soil images can be easily acquired by digital cameras, even with complicated landforms and meteorological conditions. However, few studies use this method, and further work is required to develop the ability of visible spectrum digital images to accurately estimate SWC. In this study, 42 soil samples were collected to investigate the relationship between surface GL and SWC using computer processing of soil surface images acquired by a digital camera. After establishing an equation to describe this relationship, a simple calibrated model was developed. The calibrated model was validated by an independent set of 48 soil samples. The results indicate that surface GL was sensitive to SWC. There was a negative linear relationship between surface GL and the square of SWC for the 42 calibration soil samples. Based on this negative relationship, a model was established to estimate SWC from surface GL. The results of model validation showed the estimated SWCs by surface GL were very close to the measured SWCs (P=0.01). Generally, SWC could be estimated from surface GL for a given soil, and the model could be used to quickly and accurately determining SWC from surface GL measurements.

Key words: Surface soil water content, surface gray level, soil surface image, visible spectrum