Salt tolerance of crops

One of the first studies made on soil salinity and plant response was published in the USDA Agriculture Handbook No. 60, 1954. More than 20 years later Maas and Hoffman published the results of an extensive study on salt tolerance. In 2001, a Canadian study provided a substantial amount of additional data. A comprehensive survey of tolerances reported worldwide was made by the FAO in 2002. Most studies were made with pot or drum experiments or in lysimeters under controlled conditions. The collection of field data under farmers' conditions was rare, probably due to the greater efforts and higher costs involved, the lack of control of plant growing conditions other than soil salinity, and the larger random variation in crop yields and soil salinity. Yet, with statistical methods, it is possible to detect the tolerance level from field data. Salt Farm Texel, a Dutch-based research company has identified various crops that have considerable amount of salt tolerance. ==Classification==

production and soil salinity in farmland. The salt tolerance (breakpoint, threshold) is about ECe = 3.3 dS/m Soil and water salinity can be expressed in various ways. The most common parameter used in soil salinity is the electric conductivity of the extract (ECe) of a saturated soil paste in units of deciSiemens per metre (dS/m) (previously measured in millimhos per centimeter (mmho/cm)). Bernstein presented the following soil classification based on ECe in dS/m: ECe 0–2 non-saline soil ECe 2–4 slightly saline, yield of sensitive crops reduced ECe 4–8 moderately saline, yield reduction of many crops ECe 8–16 saline, normal yield for salt tolerant crops only ECe > 16 reasonable crop yields only for very tolerant crops ==Modeling==

production and soil salinity in the field production and soil salinity in the field A common way to present crop – salinity data is according to the Maas–Hoffman model (see above figure): initially a horizontal line connected to a downward sloping line. The breakpoint is also called threshold or tolerance. For field data with random variation the tolerance level can be found with segmented regression. As the Maas-Hoffman model is fitted to the data by the method of least squares, the data at the tail-end influence the position of the breakpoint. Another method was described by Van Genuchten and Gupta. It uses an inverted S-curve as shown in the left-hand figure. This model recognizes that the tail-end may have a flatter slope than the middle part. It does not provide a sharp tolerance level. Using the Maas–Hoffman model in situations with a flat trend in the tail-end may lead to a breakpoint with a low ECe value, owing to the employment of the condition to minimize the deviations of the model values from the observed values over the entire domain (i.e. including the tail-end). Using the logistic sigmoid function for the same data applied in the van Genuchten-Gupta model, the curvature becomes more pronounced and a better fit is obtained. A third model is based on the method of partial regression, whereby one finds the longest horizontal stretch (the range of no effect) of the yield-ECe relation while beyond that stretch the yield decline sets in (figure below). With this method the trend at the tail-end plays no role. As a result, the tolerance level (breakpoint, threshold) is larger (4.9 dS/m) than according to the Maas-Hoffman model (3.3 dS/m, see the second figure above with the same data). Also a better fit is achieved. ==Augmenting tolerance==

Presently a considerable amount of research is undertaken to develop agricultural crops with a higher salt tolerance to enhance crop cultivation in salinity stricken regions. ==Leaf damage==

In Australia the following classification of sprinkler irrigation water salinity was developed: == See also ==