Citrus orchard water use, transpiration crop coefficients and orchard data, South Africa (2013-2018)

This dataset was harvested from WRC Report No. TT 772/2/18 (Volume 2 of WRC Project K5/2275), in which Citrus Research International, the University of Pretoria and the CSIR measured and modelled the water use of citrus orchards of different ages and canopy sizes. Transpiration was measured hourly with the heat ratio sap flow method in sixteen commercial orchards and upscaled to orchard level. Eight orchards were located in the winter rainfall region near Citrusdal, Western Cape (farms Patrysberg and Brakfontein: 'Midknight' and 'McLean' Valencia, 'Bahianinha', 'Washington' and 'Cambria' navel, and 'Afourer' mandarin), and eight in the summer rainfall region near Letsitele, Limpopo (Mahela Boerdery: 'Star Ruby' grapefruit, 'Midknight' Valencia and 'Valley Gold' mandarin). Measurements ran from August 2013 to March 2018, with individual orchard records of 77 to 718 days. Evapotranspiration was measured with eddy covariance during shorter campaigns in seven orchards. A water stress trial was also conducted in a 'Delta' Valencia orchard near Groblersdal in the 2016/17 season. The harvested data were compiled into a structured spreadsheet comprising an orchard register (location, planting year, rootstock, spacing, irrigation system, soil texture and measurement period); orchard characteristics (canopy dimensions, canopy cover, leaf area index, trunk circumferences, yield and water use efficiency); transpiration summaries (total, maximum, minimum, average daily and annual water use, in mm and litres per tree); eddy covariance ET, ETo and T campaign summaries with instrument heights; fortnightly transpiration crop coefficients (Kt) for fourteen orchards and proposed monthly Kt for orchards with 30, 50 and 70% canopy cover; radiation interception model inputs and model performance statistics; canopy conductance model coefficients; Kt versus canopy size regressions; and the effect of water stress on fruit mass and yield. Values were extracted from the report tables with page-level provenance and were cross-checked between tables. Inconsistencies found in the source report were flagged in a dedicated sheet rather than corrected, except where the correct value could be established with high confidence. Literature values from the review chapter were excluded. The daily and hourly time series presented only as figures in the report were not included.