Evaluation of the Watermark sensor for use with drip irrigated vegetable crops

The Watermark 200SS sensor was evaluated for the measurement of soil matric potential (SMP) with drip-irrigated vegetable crops. Pepper and melon crops were grown sequentially during autumn-winter and spring-summer, in a sandy loam soil in a greenhouse. Ranges of SMP were generated by applying three different irrigation treatments — 100, 50 and 0% of crop water requirements, during two treatment periods (16 December 2002–7 January 2003; 20 January–10 February 2003) in pepper and one treatment period (26 May–6 June 2003) in melon. Watermark sensors and tensiometers were positioned, at identical distances from irrigation emitters, at 10 cm soil depth, with four replicate sensors for each measurement. Electrical resistance from Watermark sensors and SMP from tensiometers were recorded at 30-min intervals. An in-situ calibration equation was derived using data from the first pepper treatment period. For data in the three treatment periods, SMP was calculated from Watermark electrical resistance using the in-situ, Thomson and Armstrong (in Appl Eng Agric 3:186–189 1987), Shock et al. (1998) and Allen (2000) calibration equations. Additionally, the Thomson and Armstrong (in Appl Eng Agric 3:186–189 1987) and Shock et al. (1998) equations were re-parameterised with the SOLVER® function of Microsoft Excel 2000® using data from the first pepper treatment period. Watermark-derived SMP, for each equation, were compared with tensiometer-measured SMP, for <-10, −10 to −30, −30 to −50 and −50 to −80 kPa ranges, using visual analysis, and relative root mean square error (RRMSE) and mean difference (Md) values. In rapidly drying soil, the Watermark-derived SMP responded considerably more slowly to continual drying and to drying between irrigations, regardless of the calibration equation used. Otherwise, the Watermark sensor was able to provide an accurate indication of SMP, depending on the calibration equation. The in-situ and re-parameterised equations were accurate for the conditions in which they were derived/re-parameterised. However, as the growing conditions increasingly differed from those original conditions, these equations lost their advantage compared to the two published equations, suggesting that they are not robust approaches. The Thomson and Armstrong (in Appl Eng Agric 3:186–189 1987) equation generally provided an accurate indication of SMP at >−30 kPa, measuring to −2.5 kPa. Where the soil was not drying rapidly, the Shock et al. (1998) equation generally provided an accurate indication of SMP at −30 to −80 kPa. The use of dynamic data (collected every 30 min) compared to static data (collected only at 6 a.m.) did not influence the evaluation of calibration equations. This study suggested that the Watermark sensor can provide an accurate indication of SMP provided that a suitable calibration equation is derived/verified for the specific cropping conditions, and that the performance characteristics of the sensor are considered.