*2.2.2 Crop ET coefficients [Kc = ƒ(crop phenology)]*

Actual water evapotranspirated by a crop (ETc) is not only determined by ETp; an estimation of transpiration canopy is also needed. This estimation corresponds to the concept of leaf area index (IAF; m2 of transpiring leaves/m<sup>2</sup> of cultivated land) [23, 28, 32]; for irrigation scheduling purposes, this concept is generally expressed as the "crop coefficient" (Kc) [27, 29, 33–35] which in fact is a time function, since IAF varies from bare soils at the end of winter (Kcinitial = 0.1–0.15), representing direct soil surface evaporation, up to Kcmax = 0.8–1.2), when the maximal IAF is attained. The Kc = (t) function can be represented by a double sigmoid curve (**Figure 2**) for the initial three crop phenology stages (budbreak, flowering, and veraison) [36, 37]; a constant maximal value from veraison to harvest, and a linear decline for the postharvest irrigation stage, reaching a Kcfinal = Kcinitial [37]. The maximal Kcmax value has been widely reported for most irrigated crops [27]; at flowering [Kcflower = (Kcmax/Kcinitial)/2] [26, 38].

For irrigation scheduling, the Kc daily value is obtained from **Figure 2** or from an equivalent table; the main concern is related to the onset data for each phenology stage, which seldom can be predicted accurately from crop models and needs periodic field observations throughout the irrigation season. Modifying these dates on the platform is a very simple procedure and the Kc = ƒ (t) function is easily recalculated. Different crops and/or different cultivation locations for the same crop can have quite different Kc curve (**Figure 2**) shapes, due to the relative onset date and duration of each phenology stage, but essentially, this schematic representation can be adapted to these differences.

## **2.3 Automatic Kc value adjustments, based on soil water content data**

Data on the evolution of soil water content at specific depths and distances from the irrigation lateral enable the platform to automatically adjust Kc values for the next 5 days, with the aim to increment or reduce recommendations for next irrigation dates and water depths to be applied, in order to keep a constant soil water

**Figure 2.** *The Kc = ƒ (phenology stage) function.*

availability condition; these data are obtained by using soil water content probes, providing either real-time or periodic measurements with portable soil probes. This platform feature is an independent checking for the balance between calculated ETc and actual depth water applied, enabling to automatically correct eventual errors in the calculated ETc. If the calculated ETc value is lower than the actual ETc, platform recommendations will determine underirrigation and a gradual reduction in the soil water content, while if calculated ETc > actual ETc, overirrigation will determine a gradual increment in the soil water content. Soil water content increments or reductions over 5% between consecutive measurements trigger automatic modifications on Kc values for the next 5 days and thus, the process is self-adjusted. All Kc adjustments are kept in an historical file, to be used as platform input data for the following irrigation seasons (see Section 2.2.2) [35].

Adjustment of Kc daily values related to soil water dynamics, as affected by the balance between calculated ETc and actual irrigation water depth applied, represents an automatic fine-tune procedure on irrigation scheduling, aimed to keep a constant crop water availability condition, simultaneously considering atmospheric evaporative demand, crop IAF evolution, and actual irrigation timing and water depth applied; this adjustment is seldom found in most irrigation scheduling models available in the market. We have assumed that ETp data retrieved from weather stations and actual irrigation water application are trustworthy, since modern irrigation equipment provides automatic digital operation registering options (date, time, water volume applied on each field irrigated section), including data transmission by radio frequency or through the Internet, thus reducing human intervention on data handling.
