Assessment of Basal Crop Coefficient Adjustment in Grapevines with Active Ground Cover: A Case Study

Abstract

Competition for water resources makes it necessary to advance research focused on estimating the water needs of row crops, such as vineyards. Following the FAO-56 methodology and the A&P approach, the soil water balance model was applied to a vineyard with continuous vegetation cover in temperate climate conditions (Galicia, Spain). Basal crop coefficients adjusted to local conditions were obtained for both the vineyard and the active vegetation. After SIMDualKc model adjustment, r² values greater than 0.86 were obtained, along with goodness-of-fit indicators that demonstrate the model’s ability to predict soil water content (PBIAS_avg = 1.16; EF_avg = 0.89; dIA_avg = 0.97). A correction factor is proposed that improves the partitioning of the transpiration component in row crops with active cover. The transpiration demand of the vineyard increased by 35% in four study cases (northern Portugal, northwestern Spain, and Italy). The proposed correction factor is shown to be in line with the actual conditions and complex behaviour of a vineyard with active vegetation cover, which opens the way for improved water requirement prediction in complex management situations such as the one studied here. The proposed methodology is expected to improve the efficiency of irrigation management through more accurate determination of the real water amount required by orchards.

1. Introduction

Climate conditions and water resource availability have changed very quickly over the past twenty years in many parts of the world [1]. Moreover, there is a need for more efficient agricultural production due to the increasing global population, generally concentrated in urban areas (or their influence areas), and the associated increase in demand for food. In parallel, reductions in agricultural land and limited water resources which must be shared with other uses (mainly urban and environmental) have exacerbated this problem. Water governance should be implemented holistically, considering all water uses—irrigation, urban, industrial, and environmental—at differing scales (local, regional, and global). Multiple water resources are actually available, including groundwater, surface resources, desalination, and reclaimed water; this panoply facilitates the use of a complex water mix, where water quality and availability change from day to day [2]. In this way, the agricultural sector has implemented new irrigation areas over agricultural land that were not traditionally irrigated (at nearly 20%), using drip irrigation in many cases. This irrigation systems requires less water, achieved through precise delivery at certain times and places, demonstrating clear benefits to crops [3].

Hundreds of studies over the last few years have focused on managing drip irrigation systems, irrigation scheduling, water productivity, and water use efficiency. These studies come from a perspective based on current technologies (e.g., climate forecasts, satellite data, proximal sensors) and the estimation of crop water requirements. For the latter, a well-known methodology based on FAO-56 [4] is usually implemented, where the use of crop coefficients, adjusted to local conditions and reference evapotranspiration, has yielded successful outcomes for many crop types, including field crops, orchards, vineyards, and fruit trees [5,6,7,8,9,10]. Several methodologies have been applied to implement crop coefficient databases to facilitate the determination of crop water requirements and crop evapotranspiration (ET_c). Global methods, such as soil water balance, balance energy, weighing lysimeters, and eddy covariance, have generally been successful, but there are pros and cons to all cases [5]. The FAO-56 methodology offers two approaches: simple and dual. The former does not allow for the partitioning of soil evaporation (E_s) and crop transpiration (T_crop); meanwhile, after modelling all components of water balance, the latter (Dual-Kc) approach facilitates separating ET_c into two main components: E_s and T_crop. For field crops, this separation is not generally relevant due to the short period during the growing season when there is incomplete cover, which coincides with the first stages when part of the plot is bare soil.

However, for row crops, which include fruit trees, vineyards, and blueberries, among others, there is no complete soil cover due to the crop training system, such that separation of the ET_c components is key to discerning which aspects of crop water management can be improved [10]. In row crops, the key aspect of water resource management is the possible presence or absence of active ground cover in the rows and inter-rows, which competes for available resources but offers improvements and synergies in other respects, such as erosion control, increased biodiversity, and improved soil properties.

Several studies have developed approaches to achieve better adjustments in response to real crop water requirements, mainly focusing on row crops. Allen and Pereira [11] included the density coefficient (K_d), which includes the cover fraction (f_c) and height crop values. The A&P approach was tested by Pereira et al. [6] to assess density adjustment of several crops, including vineyards. Studies have shown that similar values can be obtained when comparing the K_cb values of FAO-56 and K_cb values adjusted to density coefficient (K_d). This approach brings the obtained results closer to a real cultivation situation, and has been applied in multiple studies with good outcomes.

High-vigour row crops are generally developed in rainy regions or temperate climates, so it is common to implement a cover crop (row and inter-row) for vigour control and to maintain soil (e.g., to control erosion). This practice is linked to crop practices, traditions, and variety characteristics. Recently, it has been highlighted that European countries, through the CAP (Common Agricultural Policy), have promoted the ‘Farm to Fork’ strategy in which environmental considerations have become a key aspect of policy. One of the key topics in this regard is related to the implementation of ‘eco-schemes’ for farmers, where subsidies are granted when they adjust their agronomic practices to a special scheme [12]. Row crop farmers are encouraged to include cover crops during certain periods of the year, but the periods and other considerations differ between regions and countries. The final objective is to compensate for yield reduction when cover crops are present while, at the same time, ensuring positive environmental impacts [13]. This proposal has increased the adoption of inter-row crops by farmers, not only in traditional areas but also in arid climates. However, despite the existence of studies on cover crops (row and inter-row), greater knowledge is needed on water use in agricultural systems where they are present.

Vineyards are one of the main row crops grown worldwide, both in terms of surface area and distribution, due to the economic value of their final product: wine. Their location has been linked to rainfed agriculture in traditional areas (Europe) while, in the New World (Australia, South Africa, etc.), irrigation is normally used. This difference has been minimised with the widespread incorporation of irrigation systems in European wine-growing areas. As noted above, plant cover—encouraged by the EU (through eco-scheme subsidies) due to its positive effects (e.g., erosion control, soil conservation)—has increased significantly. Despite studies relevant to the wine sector having examined irrigation [14,15] and plant cover separately [16], modelling and detailed research on their combined effects are lacking, though a few studies exist in the literature [15,17,18,19]. Despite this, the separation of soil evaporation components and the transpiration of both crops (vineyard and plant cover) has not yet been addressed in depth.

The main objective of this study is to define a methodology to adjust the basal crop coefficients in vineyards with active ground cover, taking into account the cover fraction of the crop and active ground cover. The initial hypothesis is that the K_{cb cover} value is overestimated when modelling vineyard water requirements with the presence of active ground cover during the growing season. The key goals of the study are (1) to adequately describe active ground cover, given its influence on the soil water balance; (2) to improve the process of estimating K_{cb cover} in accordance with actual conditions; and (3) to appropriately weigh the transpiration components of the vineyard and the active ground cover separately. To reach this goal, (i) calibration and validation of the soil water balance of a vineyard with active ground cover located in temperate climate conditions (Galicia, Spain) was conducted, and (ii) a new method for adjusting K_cb to actual conditions is proposed. This facilitates the separation of crop and active ground cover transpiration (inter- and/or intra-row), so that the actual water requirements of both vineyards and cover crops can be separated correctly. The proposed methodology is expected to facilitate efficient irrigation management by determining the real water amount required by grapevines.

2. Materials and Methods

An experimental trial was conducted in a commercial vineyard of Vitis vinifera L. cv. Albariño, located in O Rosal (Galicia), owned by Bodega Lagar de Cervera, during the 2016, 2017, and 2018 growing seasons.

Specifically, the plot under study is called ‘Carballas’ by the winery due to the location of the plot, and is located at the following coordinates: 41°57′6″ N, 8°49′26″ W. It has an average altitude of 54 m and an average slope of 14.4% within the area regulated by the DO Rías Baixas. The vineyard was planted on 110-Richter in 2006, with a spacing of 2 m between plants and a separation between rows of 3 m, yielding a density of 1667 pl ha⁻¹. The vineyard is trained using a vertical lyre system with an H-shaped structure and cane pruning, with rows running east–west. Pruning is carried out in winter, leaving an average of 28 buds per vine.

The irrigation system in place on the plot was upgraded in 2016 to implement different irrigation treatments, based on variations in the water regime through the application of irrigation at different stages of the crop cycle. The irrigation system has integrated, self-compensating 2.1 L h⁻¹ drippers, separated by 0.75 m, for a dose of 0.93 L m⁻² h⁻¹.

Taking into account the growing season of the vine (April to September), the average annual ET_o [20] for the period 2011–2020 is 750 mm, slightly below the ET_o for 2016 and 2018 (761 and 758 mm) and 4.3% below the ET_o for 2017 (782 mm). This shows a trend toward increased evapotranspiration demand in recent years (Figure 1). More details about the experimental site, irrigation system, and crop conditions have been given by Fandiño [19].

For the thorough calculation of ET_o and to model crop water requirements in the study years (2016, 2017, and 2018), the FAO-56 Penman–Monteith equation [4] was applied, using the variables temperature, wind speed, hours of sunshine, and average relative humidity. The results for the three years are shown in Figure 1, together with the rainfall for each year.

The climatic conditions varied between the three years, with higher rainfall in 2016 (1844 mm) and 2018 (1500 mm), while 2017 was less rainy (969 mm). In the period from April to September, these differences between years remain, with an ET_o of 652, 655, and 642 mm and precipitation of 554, 190, and 329 mm for 2016, 2017, and 2018, respectively. This allows us to group these years together, with 2016 and 2018 as wetter years and 2017 as a dry year.

In addition to these parameters, soil hydraulic properties were determined, including the water content at field capacity (θ_FC) and permanent wilting point (θ_WP). As recommended by other authors [21,22], θ_FC and θ_WP were determined by monitoring soil water content over several years (2011–2018) using TDR equipment (TDR-100, Campbell Scientific, Logan, UT, USA). As shown in

Table 1. Soil hydraulic properties and total water available in the soils of the plot ¹.

Treatment	Sand	Clay	Silt	θFC	θWP	TAW
R0	56.5	18.5	25.0	0.280	0.095	111
R1	59.5	19.0	21.5	0.280	0.103	106
R2	58.7	18.3	23.0	0.274	0.114	96
R6	58.2	16.3	25.5	0.280	0.095	111
Average	58.2	18.0	23.8	0.279	0.102	106

Note(s): Sand, clay, and silt are recorded as a percentage. θ_FC: water content at field capacity (cm³ cm⁻³), θ_WP: water content at permanent wilting point (cm³ cm⁻³), TAW: total available water (mm). ¹ The values refer to a soil depth of 0.6 m.

, there are slight differences in the soil hydrological limits in the different treatments.

The hydrolimits presented were used to calculate the soil water balance, considering a root depth (Z_r) of 0.6 m and θsat established at 0.350 (cm³ cm⁻³).

2.1. Irrigation Treatments

The treatments implemented to evaluate different water regime strategies under study are described below. For each irrigation treatment, four repetitions (blocks) were defined, where 21 vines were controlled. In each block, two TDR probes were installed (60 cm of depth). During the three growing seasons, soil water contents were determined 10 or 11 times per season; the average values per day and treatment were used during the calibration process.

• R0. Control in rainfed conditions (0% ET_o).
• R1. Irrigation with 30% ET_o from the onset of veraison to the end of ripening.
• R2. Irrigation with 30% ET_o from the pea size phenological stage to the end of ripening.
• R6. Irrigation with 30% ET_o from the sprouting phenological stage until the end of ripening.

Regarding water management strategies, irrigation was initiated on a date characteristic of vineyard development (bud break, pea size, and/or veraison), with treatment R6 being the usual choice. Treatments R2 and R1 are similar to the usual irrigation management practices in other wine-growing regions. Irrigation was scheduled to provide, as much as possible, 30% of the weekly ET_o, forecasting it and correcting any deviations in the following week.

Table 2. Irrigation periods and total irrigation depth (mm). Years: 2016–2018.

Treatment	2016		2017		2018
	Irrigation Period	Total Depth (mm)	Irrigation Period	Total Depth (mm)	Irrigation Period	Total Depth (mm)
R0	-	-	-	-	-	-
R1	16 Aug.–01 Sep.	31.6	03 Aug.–27 Aug.	81.5	22 Aug.–14 Sep.	22.3
R2	11 Jun.–01 Sep.	130.2	15 Jun.–27 Aug.	136.6	28 Jun.–14 Sep.	73.5
R6	19 Apr.–01 Sep.	165.2	06 Apr.–27 Aug.	163.3	02 May–14 Sep.	125.6

summarises the treatments implemented in each of the three years, including the start and end dates of irrigation.

2.2. Soil Water Modelling

The water requirements of vineyards (ET_c) can be assessed with different methods and techniques; in this study, the soil water balance method was used, as previously reported [17,18,23,24].

The ET_c estimate is based on the use of the crop coefficient (K_c) as the ratio between ET_c and reference evapotranspiration (ET_o). Therefore, ET_c is obtained as ET_c = K_c × ET_o [4].

Several factors determine K_c in a vineyard, including the stage of vegetative growth, the height of the canopy (h_crop) and its architecture, the training system, the soil cover fraction, and the presence of mulch and/or active ground cover. The latter is affected by irrigation and soil management, as well as crop management and local conditions [8,10].

Since all of these factors must be considered in detail, a dual approach was selected to determine K_c [4]. The method consists of dividing K_c into two coefficients, one for crop transpiration (K_cb) and another for soil evaporation (K_e), resulting in K_c = K_cb + K_e. Therefore, actual evapotranspiration (ET_{c act}), which is lower than that obtained under standard conditions (ET_c), is defined as follows:

$${\mathrm{E}\mathrm{T}}_{\mathrm{c}\mathrm{a}\mathrm{c}\mathrm{t}}=\left({\mathrm{K}}_{\mathrm{s}}{\mathrm{K}}_{\mathrm{c}\mathrm{b}}+{\mathrm{K}}_{\mathrm{e}}\right){\mathrm{E}\mathrm{T}}_{\mathrm{o}}$$

(1)

where ET_{c act} and ET_o are expressed in mm d⁻¹; K_cb is the basal crop coefficient; K_s is the coefficient describing the effect of water stress on crop transpiration, with K_s < 1 when limitations occur due to soil water availability; and K_e is the soil evaporation coefficient.

Since the dual approach takes the presence of active vegetation and its variation throughout the cycle into account [11], its use facilitates understanding the separation of water used between crops and active ground cover, as well as soil evaporation.

In the experimental vineyard, there is permanent, active vegetation that competes with the crop for water resources; thus, it is necessary to determine its weight within the evapotranspiration processes. To use the dual approach, it is necessary to tabulate baseline crop coefficients, which must then be adjusted to local conditions. Currently, these values do not exist for vineyards with active ground cover [4,11,25], despite studies on vineyards with similar conditions. Nonetheless, these can be used as a basis for determining these values. Notably, the conditions of active ground cover vary throughout the growing season due to the operations carried out on it, as well as climatic conditions. Therefore, adjusting these conditions is complex, and the baseline coefficients must be adapted to each specific management case.

To determine the effect of active ground cover, the SIMDualKc model [26] incorporates a tool that accounts for changes in ET due to the presence of active ground cover, as well as the management of this vegetation in the vineyard. For this reason, the SIMDualKc model was selected. This model has been tested on different crops [15,27,28,29] and validated in previous studies [17,18] on Galician vineyards with active ground cover.

Therefore, the ultimate objective of modelling is to calibrate the basal coefficients at each growing stage [4], taking into account the management of active ground cover together with the estimation of soil water balance components.

This study focuses on the results related to model calibration/validation using SIMDualKc, as well as on the components of soil water balance. To determine ET_c, it is necessary to calibrate the baseline crop coefficients for each stage of crop development—a process explained in the following section. The key crop growing stages are as follows: initial (ini), rapid growth, maximum development (mid), and final (end).

The SIMDualKc model calculates the soil water balance in the root zone, expressed in terms of water depletion at the end of the day, using the following layer model [4,25]:

$${\mathrm{D}}_{\mathrm{r},\mathrm{i}}={\mathrm{D}}_{\mathrm{r},\mathrm{i}-1}-{\left(\mathrm{P}-\mathrm{R}\mathrm{O}\right)}_{\mathrm{i}}-{\mathrm{I}}_{\mathrm{i}}-{\mathrm{C}\mathrm{R}}_{\mathrm{i}}+{\mathrm{E}\mathrm{T}}_{\mathrm{c},\mathrm{i}}+{\mathrm{D}\mathrm{P}}_{\mathrm{i}}$$

(2)

where D_r,i is the soil water depletion in the root zone at the end of day i (mm); D_r,i−1 is the soil water depletion in the root zone at the end of day i − 1 (mm); P_i is the precipitation on day i (mm); RO_i is the runoff from the soil surface on day i (mm); I_i is the net amount of irrigation on day i that infiltrates the soil (mm); CR_i is the capillary rise from the water table on day i (mm); and DP_i is the water flow leaving the root zone by deep percolation on day i (mm). To achieve root zone balance, it is necessary to first adjust the balance in the surface layer of the soil, which is directly involved in soil evaporation processes (E_s) and directly related to moisture and the fraction of soil cover by crops and active ground cover.

SIMDualKc Model

A detailed description of the SIMDualKc model has been provided by Rosa et al. [26], who set out the data required for the model to function and the method for calculating the soil water balance. This covers both the surface layer and the root zone, as well as the different extensions it incorporates to determine surface runoff, deep percolation, and the effects of active ground cover management on crop water requirements.

To calculate the water balance in the surface layer, the percentages of sand and clay are required to determine the readily evaporable water (REW) (mm); in addition, the soil water content of the surface layer (cm³ cm⁻³) at field capacity (θ_FC = 0.254 cm³ cm⁻³) and the permanent wilting point (θ_WP = 0.089 cm³ cm⁻³) are required, together with the depth of the soil surface layer subject to evaporation (Z_e = 0.15 m). The latter facilitates the determination of total evaporable water (TEW, mm).

Using data from a nearby agrometeorological station (MeteoGalicia, Xunta de Galicia), ET_o (mm) was calculated using the Penman–Monteith equation with limited data. We then introduced P (mm) and, finally, RH_min (%) and u₂ (m s⁻¹), which are required to correct crop coefficients to local conditions [4]. Irrigation scheduling and depths were introduced to simulate the soil water balance with all parameters (Table 2), taking into account the wetting fraction of the soil caused by the irrigation system (f_w = 0.10).

Moreover, information about crop growth stages (dates, height, and fraction of ground cover) is required. In addition, it is necessary to enter the values for the soil water depletion fraction (p) for each development stage (p_ini, p_mid, and p_end), which are used to determine the readily available water (RAW) for the crop (mm). It is important to enter the crop’s basal cultural coefficients (K_{cb full}) when there is total soil coverage without water restrictions. This is related to the maximum height of the crop (h) in each phase (K_{cb full ini}, K_{cb full mid}, and K_{cb full end}) (

Table 3. Crop growth stages, height (m), and fraction of ground cover (f_c).

Stages	2016			2017			2018
	Date	h (m)	fc	Date	h (m)	fc	Date	h (m)	fc
Start	09 Mar.	1.0	0.01	01 Mar.	1.0	0.01	16 Mar.	1.0	0.01
Start rapid grow	07 Apr.	1.3	0.05	20 Mar.	1.3	0.05	04 Apr.	1.3	0.05
Start maximum grow	12 Jun.	2.0	0.20	16 May.	2.0	0.20	11 Jun.	2.0	0.20
Start maturation	16 Aug.	2.1	0.25	03 Aug.	2.1	0.25	16 Aug.	2.1	0.25
Harvest	26 Sep.	1.9	0.20	07 Sep.	1.9	0.20	20 Sep.	1.9	0.20

).

As grapes are a row crop, it is necessary to enter the required parameters to adjust the K_cb to the current density conditions (K_d), orientation (east–west), and row width (width, m) of the crop; the latter are viewed from the east–west direction. Finally, we incorporate the value of K_{c min} (0.15), the minimum value of the crop coefficient (dry soil without vegetation).

Given that there is active ground cover in the vineyard under study, when applying the A&P approach [11] to adjust the basal crop coefficients (K_cb) to the crop height (h) and its density, it is necessary to determine the basal crop coefficient during the maximum development stage for vegetation with complete ground cover (K_{cb full}).

$${\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{f}\mathrm{u}\mathrm{l}\mathrm{l}}={\mathrm{F}}_{\mathrm{r}}\left(\mathrm{m}\mathrm{i}\mathrm{n}\left(1.0+0.1\mathrm{h};1.2\right)+\left[0.04\left({\mathrm{u}}_2-2\right)-0.004\left({\mathrm{R}\mathrm{H}}_{\mathrm{m}\mathrm{i}\mathrm{n}}-45\right)\right]{\left({\displaystyle \frac{\mathrm{h}}{3}}\right)}^{0.3}\right)$$

(3)

where h, u₂, and RH_min are as previously defined, and F_r is the stomatal resistance correction factor [4,6].

The K_{cb full} value represents an upper general limit for K_{cb mid} when there is full soil cover without water restrictions. This expression suggests that the upper limit for K_{cb full} is 1.20 plus adjustments for local climatic conditions.

Since there is active ground cover in the vineyard, to determine K_cb—mainly representing transpiration—it is necessary to consider the transpiration of active ground cover (K_{cb cover}), which can be expressed as a function of density (K_d) [6,11] as follows:

$${\mathrm{K}}_{\mathrm{c}\mathrm{b}}={\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}+{\mathrm{K}}_{\mathrm{d}}\left(\max \left[{\mathrm{K}}_{\mathrm{cb}\mathrm{full}}-{\mathrm{K}}_{\mathrm{cb}\mathrm{cover}},{\displaystyle \frac{{\mathrm{K}}_{\mathrm{cb}\mathrm{full}}-{\mathrm{K}}_{\mathrm{cb}\mathrm{cover}}}{2}}\right]\right)$$

(4)

where K_cb is determined by considering the structure of the vegetation and its extent, through the density coefficient (K_d) defined by Allen and Pereira [11], including the transpiration of active ground cover and taking into account K_{cb cover}.

To determine the other parameters involved in soil water balance, capillary rise, runoff, and deep percolation, methodologies proposed by other authors were applied [25,26]. Capillary rise was considered negligible since the water table is very far from the surface. For deep percolation (DP), a parametric function defined by Liu et al. [30] was used. The required parameters are soil water content at saturation (θ_SAT, cm³ cm⁻³), field capacity (θ_FC), and wilting point (θ_WP), as well as the parameters a and b, adjusted to soil conditions.

To estimate runoff (RO), the curve number (CN) method developed by the USDA-SCS [31] was used, based on the CN and water depletion in the evaporable layer (D_e) of the soil. Allen et al. [25] have provided tabulated CN values according to soil type and crop; accordingly, a CN value of 64 was adopted in the modelling process, intermediate to coarse and medium texture soil.

2.3. Active Ground Cover—Modelling

To determine the combined K_cb of active ground cover and crops (Equation (4)), referred to in the results as K_{cb cover+crop}, the extension included in the SIMDualKc model [26] is used. To use this extension, it is necessary to enter the initial and end values: the fraction of soil cover by active ground cover (f_{c cover}), the density and height (h_cover). Moreover, the update of active ground cover conditions throughout the cycle, as different cultural operations are carried out, is required (

Table A1. Density and height of active ground cover. Years: 2016–2018.

Year	Date	Day of Year	Row		Inter-Row
			Density	h (m)	Density	h (m)
2016	09 Mar.	69	0	0	0.80	0.05
	24 Mar.	84	0	0	0.80	0.05
	04 May	125	0	0	0.80	0.15
	16 May	137	0.20	0.05	0.70	0.15
	12 Jun.	164	0	0	0.70	0.15
	26 Jun.	178	0	0	0.80	0.20
	17 Jul.	199	0	0	0.70	0.15
	30 Aug.	243	0	0	0.60	0.10
	18 Sep.	262	0	0	0.50	0.05
	26 Sep.	270	0	0	0.50	0.05
2017	01 Mar.	60	0	0	0.80	0.15
	16 May	136	0	0	0.80	0.15
	31 May	151	0.20	0.05	0.70	0.10
	15 Jun.	166	0.40	0.10	0.65	0.10
	05 Jul.	186	0	0	0.50	0.10
	19 Aug.	231	0	0	0.40	0.05
	30 Aug.	242	0	0	0.40	0.05
	07 Sep.	250	0	0	0.40	0.05
2018	16 Mar.	75	0.20	0.05	0.80	0.20
	05 Apr.	95	0.20	0.05	0.80	0.25
	16 May	136	0.15	0.05	0.70	0.15
	11 Jun.	162	0	0	0.60	0.10
	16 Aug.	228	0	0	0.40	0.05
	20 Sep.	263	0	0	0.45	0.05

).

The procedure used to determine K_{cb cover} consists of using the concept of K_{cb cover full}, which depends on vegetation height and local climatic conditions (RH_min and u₂), adjusted to daily changes in active ground cover density using the density coefficient (K_{d cover}). The model uses a variant of Equation (4) [6,25], considering a K_{c min} of 0.15, which is the minimum crop coefficient for bare soil (Equations (5)–(8)).

$${\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{f}\mathrm{u}\mathrm{l}\mathrm{l}}=\mathrm{m}\mathrm{i}\mathrm{n}(1.0+0.1{\mathrm{h}}_{\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}};1.2)+\left[0.04\left({\mathrm{u}}_2-2\right)-0.004\left({\mathrm{R}\mathrm{H}}_{\mathrm{m}\mathrm{i}\mathrm{n}}-45\right)\right]{\left({\displaystyle \frac{{\mathrm{h}}_{\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}}{3}}\right)}^{0.3}$$

(5)

$${\mathrm{f}}_{\mathrm{c}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}=\mathrm{d}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{d}.{\mathrm{f}}_{\mathrm{r}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}$$

(6)

$${\mathrm{K}}_{\mathrm{d}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}=\mathrm{m}\mathrm{i}\mathrm{n}\left(1,{\mathrm{M}}_{\mathrm{L}}{\mathrm{f}}_{\mathrm{c}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}},{{\mathrm{f}}_{\mathrm{c}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}}^{{\displaystyle \frac{1}{1+{\mathrm{h}}_{\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}}}}\right)$$

(7)

$${\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}={\mathrm{K}}_{\mathrm{c}\mathrm{m}\mathrm{i}\mathrm{n}}+{\mathrm{K}}_{\mathrm{d}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}\left({\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{f}\mathrm{u}\mathrm{l}\mathrm{l}}-{\mathrm{K}}_{\mathrm{c}\mathrm{m}\mathrm{i}\mathrm{n}}\right)$$

(8)

where f_{c eff cover} refers to the effective f_{c cover}, and M_L is a multiplier on f_{c eff cover} describing the effect of canopy density on shading and on maximum relative ET per fraction of shaded ground [1.0–2.0].

As with the calculation of K_{cb full crop}, it is necessary to define the value of F_r for the active ground cover, as well as its maximum height (h_{cover max}). For the mixture of spontaneous vegetation in the vineyard studied, F_r = 0.55 and h_{cover max} = 0.3 (m) were considered.

K_{cb cover} Adjustment to Local Conditions

As the K_{cb cover} provided by the SIMDualKc model is determined in the absence of crops (vineyards), it is necessary to estimate this coefficient adjusted to crop development conditions. Although these conditions (i.e., the absence of a crop) can be assumed in the initial phase of the cycle, in the other phases, this assumption would be made to the detriment of the crop’s own transpiration, as the transpiration of the active ground cover would increase. Thus, it is necessary to define a new coefficient that corrects K_{cb cover}, which should reflect the actual situation of active ground cover and separate the vineyard transpiration and active ground cover transpiration. To this end, a sensitivity analysis of the new coefficient was carried out, considering scenarios with different coverage fractions (f_c) common in vineyards [10].

Moreover, to evaluate the performance of this new factor, previous articles describing the information needed to validate them were consulted. All studies used soil water balance and the Dual-Kc approach in row crops with active ground cover. The necessary values are K_{cb cover}, f_{c crop}, and h_crop, applying the new correction factor for the different growth periods of the vineyard.

2.4. Statistical Analysis

To minimise the differences between the observed soil water content (SWC_obs) and the simulated soil water content (SWC_sim), the calibration process involved adjusting the crop parameters K_cb and p, related to the initial, peak development, and final stages; the soil evaporation parameters TEW, REW, and Ze; and the DP parameters, a, and b. Data from 2018 were used in the calibration process for treatment R0. A set of initial parameters related to crop parameters (K_cb and p) was selected [4,25]; the parameters for soil water evaporation were taken from soil water content measurements in the surface layer from previous projects; and the DP parameters were taken from information provided by Liu et al. [30]. The initial soil water conditions in 2018 for treatment R0 showed a reduction value of 0% for TEW and TAW.

The trial-and-error process of calibrating the parameters begins with the K_cb values. In the following phases, the trial-and-error process is also applied to the parameter p and then to the aforementioned E_s and DP parameters. Once the process is complete, the calibrated parameters are obtained, which must then be validated in the remaining treatments and years [26]. The calibration–validation process is considered satisfactory when the goodness-of-fit indicators related to validation varies by less than 20% from those related to calibration.

To validate the accuracy and goodness of fit of the model predictions, two approaches were taken. The first consisted of a graphical analysis of the annual evolution of the values simulated by the model compared to the observed water contents. This first approach provides an adequate perception of the existence of trends or biases in the modelling (if any). The second approach involved performing a regression analysis, forced to the origin, between the observed data and those simulated by the model. When the regression coefficient (b) is close to 1.0, the covariance is close to the variance of the observed values, indicating that the simulated and observed values are statistically similar. In addition, a coefficient of determination (r²) close to 1.0 indicates that most of the total variance in the observed values is explained by the model.

Finally, a set of indicators of estimation errors in the residuals is determined, as reported in previous work [32,33] and described in detail by Pereira et al. [34]. In addition, to assess any bias in the trend of the estimates, the percentage bias (PBIAS) (%) was determined. PBIAS measures the average tendency of predictions to be larger or smaller than their corresponding observations, where positive values indicate an underestimation bias and negative values indicate an overestimation bias. The indicators determined were the normalised root mean square error (NRMSE) (%); the root mean square error (RMSE); the mean absolute error (MAE); the mean relative error (ARE); and the maximum error (Emax), in mm. In addition to determining various model fit quality indicators—such as the fit index (d_IA) [35], which represents the relationship between the mean square error and the potential error due to modelling; and, finally, the model efficiency (EF)—a normalised statistic developed by Nash and Sutcliffe [36] was used to determine the relative magnitude of the residual variance compared with the variance in the observed data.

3. Results

3.1. Soil Water Balance—Calibration and Validation

The SIMDualKc model was calibrated using soil water content (SWC, mm) data determined in the field for treatment R0 in 2018. It was also validated for the remaining treatments and years. The standard and calibrated crop parameters are shown in

Table 4. Standard and calibrated modelling parameters.

Parameters	Standard	Calibrated
Kcb full ini	0.20 1	0.33
Kcb full mid	0.80 1	0.64
Kcb full end	0.60 1	0.48
pini	0.45 2	0.60
pmid	0.45 2	0.50
pend	0.45 2	0.60
a	305 3	290
b	−0.0173 3	−0.0320

Note(s): ¹ [11]; ² [4]; ³ [30].

. It should be noted that the standard crop coefficients refer to vineyards without active vegetation cover. The parameter p do not require major adjustments with respect to p = 0.45 [4,25], so an adjustment was simply made to local climatic conditions, according to Allen et al. [4]. M_L values are expected to vary between 1.5 and 2.0 [25], depending on the density and thickness of the vegetation canopy. For a theoretical vineyard with partial coverage, as a result of surface roughness and due to the high amount of energy available resulting from the open space between crop rows [37], M_L = 1.5 is considered a reasonable value under the study conditions.

Figure 2, Figure 3 and Figure 4 show the graphical results of the model adjustment between the observed and simulated soil water content (SWC, mm) values throughout the season for the three study years and the four irrigation treatments considered. The simulated SWC values are consistent with the values observed during the three campaigns, despite the different climatic conditions between them (Figure 1) and the different water regimes applied. In 2017, during the peak development phase, differences were observed between the simulated and observed values for R1 and R2 due to errors in the management of the irrigation system, which resulted in adjustment indicators below the average obtained for the remaining treatments and years of study.

The adjustment of simulated values against observed values in 2016 (Figure 2) shows that the model tends to underestimate SWC for treatments R0 and R1, while there is a slight overestimation of observed SWC values for R6 (

Table 5. Linear regression coefficients and goodness-of-fit indicators related to the calibration and validation of the SIMDualKc model for different irrigation treatments.

Treat.	Linear Relation		Residues
	b	r2	NRMSE (%)	PBIAS (%)	RMSE (mm)	dIA	EF	Emax (mm)	AAE (mm)	ARE (mm)
2016
R0	0.98	0.98	5.09	2.20	6.28	0.99	0.96	12.20	5.07	4.62
R1	0.98	1.00	4.02	2.55	5.53	0.99	0.97	9.92	4.28	3.99
R2	1.00	0.95	3.42	−0.32	4.41	0.98	0.92	8.26	3.44	2.73
R6	1.03	0.93	4.38	−2.81	6.52	0.97	0.88	13.67	5.22	3.72
2017
R0	0.98	0.97	6.47	3.17	7.27	0.99	0.94	10.41	6.87	6.80
R1	0.97	0.86	9.41	3.68	12.25	0.95	0.77	33.80	8.81	7.07
R2	0.98	0.64	6.96	2.03	8.84	0.89	0.58	20.14	6.34	5.10
R6	1.00	0.86	3.90	−0.51	5.83	0.96	0.86	9.96	4.75	3.32
2018
R0	1.01	0.99	3.18	−0.54	3.85	1.00	0.99	5.64	3.51	3.11
R1	0.98	0.97	4.89	1.96	6.64	0.99	0.96	12.91	5.37	4.40
R2	0.98	0.94	5.75	1.50	7.11	0.98	0.93	14.60	5.52	4.83
R6	0.99	0.96	3.72	1.02	5.61	0.99	0.95	12.19	4.46	3.20
min.	0.97	0.64	3.18	−2.81	3.85	0.89	0.58	5.64	3.44	2.73
max.	1.03	1.00	9.41	3.68	12.25	1.00	0.99	33.80	8.81	7.07
Average	0.99	0.92	5.10	1.16	6.68	0.97	0.89	13.64	5.30	4.41

). The maximum differences occur at the maximum development stage, related to irrigation and precipitation events (Figure 2).

For 2017, the adjustment of simulated values against observed values shows a tendency for the model to slightly underestimate SWC for R0, R1, and R2 (b < 1.00, Table 5). Once again, the greatest differences occur at the maximum development stage, especially in treatments R1 and R2 (Figure 3), mainly due to the irrigation management problems mentioned above.

The model’s tendency in 2018 is to underestimate SWC values in irrigated treatments (R1, R2, and R6), despite the fact that the problems related to the irrigation system were solved for the current year. This can be seen at the maximum development stage where, despite there being differences, these values are minimised compared with previous campaigns (Figure 4).

3.2. Soil Water Balance—Model Fitting

The model fit was evaluated using goodness-of-fit indicators, which are presented in Table 5, together with the linear regression coefficients. The results show that the coefficients of determination (r²) vary between 0.64 and 1.00 for all treatments, with an average value of 0.92. As the average regression coefficient (b) is 0.99, in general terms, the model underestimates SWC, except in 2016 for R6 and in 2018 for R0. In contrast, the PBIAS indicator shows positive values, which means that the model slightly underestimates SWC values, although these PBIAS values do not exceed 4%. In the case of R2 and R6 in 2016, R6 in 2017, and R0 in 2018, the PBIAS is negative, indicating a slight overestimation. These values are consistent with those obtained by other authors [15,29].

Figure 5 shows the results of the linear regression when all simulated and observed SWC data are considered, verifying the results discussed above. This indicates that the values were statistically similar for all treatments and years.

The residual estimation errors (Table 5) show that the NRMSE varies between 3.18 and 9.41%, with average values around 5%, while the RMSEs range from 3.85 to 12.25 (mm), with the maximum values corresponding to treatments R1 and R2 in 2017. The average RMSE for all years and treatments is 7 mm, which represents 6.3% of the total available water in the soil (TAW); these are relatively low values. The AAE values vary between 3.44 and 8.81 (mm), which are also low values with an average representing 5% of the TAW, in line with the ARE values. Finally, the average value of E_max is 13.64 mm, generally related to the occurrence of rainfall or irrigation events.

The efficiency index (EF) was greater than 0.77, indicating that the residual variance due to modelling is comparable to the variance of the measured data (Table 5), with the exception of R2 in 2017, where the EF is lower. The lower EF index with respect to R2-2017 was due to a slight variation in the soil water content during the growing season (Figure 3), linked to operational mistakes in using the irrigation systems. Furthermore, in that year, a higher irrigation depth than planned was applied in the R2 treatment at the end of the season. Nevertheless, this specific result does not affect the capability of the SIMDualKc model to predict soil water content and use it to partition the crop and active ground cover transpiration, as denoted in the rest of the validation cases (Table 5). Finally, d_IA varies between 0.89 and 1.00, with an average value of 0.97, indicating the good fit of the model after calibration.

3.3. New Approach: Kcb Cover Adjustment

The K_cb of active ground cover (K_{cb cover}) was not obtained through calibration–validation from standard values; instead, it was calculated with the model using the procedure referred to in Section 2.3, based on observations of the fraction of active ground cover (f_{c cover}), its density, and its height (h_cover) throughout the measurement campaign (Table A1). As this K_{cb cover} value increases the transpiration of active ground cover, a new coefficient (K_{cb cover est}) has been developed to reduce it, which corrects the K_{cb cover} obtained with the SIMDualKc model, adjusting it to the vegetation of the vineyard (Equation (9)).

$${\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{t}}={\mathrm{K}}_{\mathrm{c}\mathrm{b}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}}\left[\mathrm{m}\mathrm{a}\mathrm{x}\left(1-{\mathrm{f}}_{\mathrm{c}\mathrm{c}\mathrm{r}\mathrm{o}\mathrm{p}};0.5\right)\right]$$

(9)

where f_{c crop} is the vineyard coverage fraction throughout the cycle, and 0.5 is a limiting factor for the reduction in active ground cover transpiration (T_cover). A linear relationship is selected, as this is the same model applied to vegetative growth in the SIMDualKc model. Equation (9) involves only the vineyard canopy cover fraction (f_{c crop}), thereby simplifying its application. f_{c crop} is calculated by the model in accordance with the parameters entered into the modelling for each vegetative stage.

To develop this correction factor, a trial-and-error procedure was followed in which the parameters involved in the equation were adjusted. Different scenarios have been proposed in terms of crop development, particularly regarding the maximum coverage fraction (f_{c crop}). The sensitivity analysis (Figure 6) shows a linear reduction as f_{c crop} increases, up to a maximum correction factor of 0.5, as reflected in Equation (9). The results in Figure 6 are obtained from the K_{cb cover} values simulated with the SIMDualKc model for the three study years. The coverage fraction was determined in the field for the vineyard studied (f_{c crop} = 0.25, low density), compared with other common coverages in the vineyard, at medium and high density: f_{c crop} = 0.50 and f_{c crop} = 0.70, respectively [11].

Regarding the behaviour of K_{cb cover est} for different years and densities, it tended to decrease in the initial phase of crop development, while remaining stable in the maximum vineyard development phase. These values are adapted to the development of the vineyard and, therefore, to its transpiration demand as the plant canopy increases during the initial and rapid growth phase. The lower K_{cb cover est} values are above 0.10–0.15 over the three study years (Figure 6), reflecting the minimum K_{cb cover} value usually used for crops (K_{c min}).

Active ground cover throughout the crop cycle contributes to the transpiration of the vineyard and active ground cover as a whole (K_{cb cover+crop}), with lower K_{cb cover} values in the final stages of the crop, coinciding with maximum vegetative development and the highest fraction of vineyard cover.

3.4. Crop Coefficient Partitioning

The variation in crop coefficients throughout the three seasons is shown in

Table 6. Average values of crop coefficients for the different crop stages.

2016
Stages	Kcb cover	Kcb crop	Kcb (cover+crop) act				Ke				Kc act
			R0	R1	R2	R6	R0	R1	R2	R6	R0	R1	R2	R6
Initial	0.25	0.01	0.26	0.26	0.26	0.26	0.86	0.86	0.86	0.86	1.12	1.12	1.12	1.12
Rapid growth	0.22	0.14	0.36	0.36	0.36	0.36	0.60	0.60	0.60	0.62	0.96	0.96	0.96	0.98
Max. Develop.	0.19	0.28	0.30	0.32	0.44	0.48	0.08	0.08	0.14	0.15	0.37	0.40	0.58	0.62
Harvest	0.17	0.25	0.22	0.32	0.38	0.43	0.19	0.22	0.22	0.22	0.42	0.54	0.60	0.65
Average	0.21	0.19	0.30	0.32	0.37	0.40	0.39	0.39	0.41	0.42	0.68	0.71	0.79	0.81
2017
Stages	Kcb cover	Kcb crop	Kcb (cover+crop) act				Ke				Kc act
			R0	R1	R2	R6	R0	R1	R2	R6	R0	R1	R2	R6
Initial	0.27	0.01	0.28	0.28	0.28	0.28	0.78	0.78	0.78	0.78	1.06	1.06	1.06	1.06
Rapid growth	0.24	0.12	0.36	0.36	0.36	0.36	0.47	0.48	0.47	0.50	0.84	0.84	0.83	0.86
Max. Develop.	0.19	0.27	0.36	0.38	0.45	0.46	0.11	0.11	0.15	0.17	0.47	0.49	0.60	0.63
Harvest	0.16	0.26	0.10	0.37	0.42	0.42	0.03	0.09	0.09	0.09	0.13	0.46	0.51	0.52
Average	0.21	0.20	0.30	0.36	0.40	0.41	0.27	0.28	0.30	0.31	0.58	0.65	0.70	0.72
2018
Stages	Kcb cover	Kcb crop	Kcb (cover+crop) act				Ke				Kc act
			R0	R1	R2	R6	R0	R1	R2	R6	R0	R1	R2	R6
Initial	0.27	0.01	0.28	0.28	0.28	0.28	0.87	0.87	0.87	0.87	1.15	1.15	1.15	1.15
Rapid growth	0.23	0.12	0.35	0.35	0.35	0.35	0.49	0.50	0.49	0.52	0.85	0.85	0.84	0.87
Max. Develop.	0.17	0.27	0.35	0.38	0.37	0.45	0.10	0.10	0.14	0.16	0.45	0.48	0.51	0.60
Harvest	0.16	0.25	0.15	0.22	0.27	0.39	0.06	0.11	0.12	0.12	0.21	0.33	0.40	0.51
Average	0.20	0.19	0.31	0.33	0.34	0.39	0.31	0.32	0.34	0.35	0.62	0.65	0.67	0.74
Average 2016–2018	0.20	0.19	0.30	0.34	0.37	0.40	0.32	0.33	0.35	0.36	0.63	0.67	0.72	0.76

Note(s): K_{cb cover}: basal coefficient of active ground cover using Equation (9); K_{cb crop}: basal crop coefficient; K_{cb (cover+crop) act}: basal crop coefficient of vineyard and active ground cover adjusted; K_e: soil evaporation coefficient; K_{c act}: simple crop coefficient adjusted.

for K_{cb cover}, K_{cb crop}, K_{cb (cover+crop) act}, K_e, and K_{c act}. The K_e values decrease from the initial phase (K_e = 0.84)—when there is no shading from the plant canopy, coinciding with the highest periods of rainfall—to the final phase (K_e = 0.13), when the plant canopy is fully developed (Table 6).

The initial values of K_{cb (cover+crop) act} are relatively low (0.27) due to the initial lack of vegetative development in the vineyard, together with the presence of active ground cover in the soil with moderate vigour, resulting in a K_{cb cover} of 0.26. During the rapid growth period, K_{cb (cover+crop) act} increases from 0.27 at the end of the initial period to 0.46 at the beginning of the maximum development phase. This is mainly due to crop development, where K_{cb crop} reaches maximum values (Table 6), and not to the presence of active ground cover. During this period, active ground cover undergoes different cultural operations (herbicide, mowing, etc.) more intensively than in the initial stages, so K_{cb cover} is reduced compared to the values in the initial phase.

A reduction in K_{cb (cover+crop) act} is observed in treatments R0, R1, and R2, where the potential values are not reached (Table 6). In the final phase of the crop cycle, a K_{cb (cover+crop) act} close to the potential was observed for treatment R6 (0.41), while in the remaining treatments, there is a considerable reduction in K_{cb (cover+crop) act} compared with the potentials of 0.16, 0.30, and 0.36 for R0, R1, and R2, respectively, which is consistent with the water regime applied.

4. Discussion

4.1. Soil Water Balance—Dual-Kc Approach

Over the three years, there was high variability in the SWC measurements, with the error bars showing soil heterogeneity within the same treatment, despite the consistent distance between the TDR measurement points in all cases. This variability is greater in irrigated treatments due to the arrangement of the integrated drippers in the line, as, being 0.75 m apart, the measurement points may be more or less influenced by irrigation. The slope of the plot must also be considered, as the TDR measurement points located downstream are influenced by irrigation from the upper part of the plot. The slope of the plot meant that, in some areas, the irrigation depth was not exactly the same due to changes in pressure and, thus, in opening and closing irrigation times; this resulted in intra-plot irrigation variability (Figure 2, Figure 3 and Figure 4). By arranging the blocks (replicates) per treatment at different elevations, the uncertainty of the results was reduced, ensuring the representativeness and reliability of the model for the plot as a whole, the grape variety, and the vineyard management conditions.

Under Galician conditions, ET_c is less than 5 mm day⁻¹ for most of the growing season, resulting in p-values higher than the reference values (p = 0.45) (Table 4). During the mid-season, the values were close to the reference values to p due to higher ET_c. It should be noted that the ET_c used for the numerical estimation of p represents the combined effect of the vineyard and active ground cover. The value of p for a vineyard with active ground cover should be lower, as proposed by some authors [16], based on competition for available soil water between the vineyard and the active ground cover. However, under similar temperate climate conditions (northern Portugal), other authors have obtained similar p-values [15], clearly indicating that climatic conditions are the key factor in determining the parameter p in each case.

The initial values of K_{cb full} for the initial, maximum development, and final phases were taken from Allen et al. [25] and Allen and Pereira [11] for winemaking vineyards, and were adjusted during calibration. The calibrated K_{cb full} values (K_{cb full ini} = 0.33, K_{cb full mid} = 0.64, and K_{cb full end} = 0.48) (Table 4) were higher than the standards for K_{cb full ini}, while those for the maximum development and final phases were slightly lower. The K_{cb full} values are lower than those reported by previous authors [14] for vineyards with bare soil, who obtained a constant K_{cb full} value of 0.78–0.79 in all phases. When selecting the standard starting coefficients, Allen and Pereira [11] included an implicit water-stress coefficient (K_s), contrary to K_{cb full} values tabulated for wine grapes [25]; in both cases, the presence of active ground cover is not considered, unlike in the present study. The calibrated values for K_{cb full mid} and K_{cb full end} are lower than those proposed by previous authors [25] because the f_{c eff} in the study vineyard (f_{c eff} = 0.25) is lower than that they considered (f_{c eff} = 0.5). For K_{cb full ini}, the calibrated values were higher due to the presence of active ground cover; this value is similar to the K_{cb full ini} for other fruit crops, such as apple, avocado, apricot, pear, and peach [11,25].

If we compare the average K_{cb (cover+crop) act} values (Table 6) for R6 (K_s = 1.0), referring to the active ground cover and vineyard as a whole, with the values proposed by Allen and Pereira [11] for low densities (f_{c eff} = 0.25)—implying a stress coefficient (K_s = 0.7)—we obtain K_{cb mid} = 0.46 and K_{cb end} = 0.41. These values are lower than those reported previously [11] (K_{cb mid} = 0.57 and K_{cb end} = 0.43), due to differences in planting density, age, and vineyard training systems.

4.2. Kcb Cover Adjust—New Approach

The K_{cb cover} correction factor must be calibrated in the field by measuring the transpiration of the vineyard and/or active ground cover to validate the results obtained with Equation (9). In this study, K_{cb cover} was approximated empirically by visually observing the active ground cover and vineyard (f_{c crop}; f_{c cover}), and determining the canopy conditions.

The average K_{cb cover} values for each cultivation stage are shown in Table 6 for all treatments and years of study. These values remained stable at around 0.20, with slight reductions in the maximum development and final stages. A similar behaviour of K_{cb cover} was obtained for the Albariño variety, grown on semi-trellises, with f_{c crop} = 0.55 [17]; for the Loureiro variety, with f_{c crop} = 0.40 [15]; and for Italian vineyards (Barbera), grown on vertical-shoot-positioned trellises, with f_{c crop} = 0.27 [16].

These values are slightly lower than those obtained for the Loureiro variety [15] (K_{cb cover} = 0.23), albeit well below the results for the Albariño variety (K_{cb cover} = 0.36) [17] and identical to those of Italian vineyards [16] regarding the case study with active ground cover. Notably, all previous studies [15,16,17] used a K_{cb cover} value simulated with the SIMDualKc model, which is why the values reported by the aforementioned authors are higher than those obtained in the present study.

The average K_e for the three seasons and all growing stages was 0.34, higher than that reported for the Albariño variety (K_e = 0.09) [17], grown on semi-trellises, and lower than that reported for the Loureiro variety in northern Portugal (K_e = 0.41) [15].

The K_{cb crop ini} value was lower than that obtained by previous authors [14,17]; similarly, in the maximum development and final stages, the values were lower than those reported by these authors. López-Urrea et al. [14] obtained a K_{cb crop mid} of 0.41 and a K_{cb crop end} of 0.61, higher than the respective values of 0.27 and 0.25 obtained in this study. These values are lower due to the presence of active ground cover in the case of Albariño; if we compare the K_{cb crop} values for the Tempranillo variety [14] with the K_{cb (cover+crop) act} for Albariño, we obtain similar values in the maximum development phase (0.46) and lower values for the final phase (0.42). However, Darouich et al. [16] obtained similar K_{cb crop mid} (0.28) and K_{cb crop end} (0.18) values compared with those in our study. A study conducted at a Galician vineyard [18] obtained identical values for the Godello variety grown on trellises, with active ground cover that was transformed into mulch after the initial phase of cultivation.

Analysing the three years together, and taking into account the variability of the results according to the irrigation treatment, we find that in the initial phase, the value of K_{cb (cover+crop) act} is 0.27 in all water regimes. This is lower than values obtained in previous studies [17,38], although it is in line with those obtained in northern Portugal [15]. For the maximum development phases, a K_{cb (cover+crop) act} value of 0.46 is obtained for R6 and between 0.30 and 0.45 for treatments R0, R1, and R2. In the case of the Loureiro variety, ref. [15] reported values between 0.27 and 0.48, depending on the water regime applied and the climatic conditions of the year, similar to those obtained in this study. In contrast, in a previous study under temperate climate conditions, ref. [17] reported average values for the maximum development phase of 0.65—higher than those obtained in the current study—mainly due to the training system evaluated, with semi-trellises used (as opposed to trellises), even though they used the same variety, Albariño. Finally, in the final stage, the average K_{cb (cover+crop) act} is 0.41 for R6 and between 0.10 and 0.42 for the remaining treatments. For R6, values higher than those reported by Fandiño et al. [17] (0.30) were obtained, albeit close to those obtained in R1. The “full irrigation” treatment implemented on the Loureiro variety in Portugal [15] shows a K_{cb (cover+crop) act} value of 0.41, equal to that obtained for R6; however, the value for rainfed land was 0.23, higher than R0 (K_{cb (cover+crop) act} = 0.16) in the present study.

4.3. Validation of Kcb Cover Adjust Study Cases

Studies of soil water balance in vineyards with active ground cover remain limited [15,16,17,18,19]; however, all authors have achieved good results after K_cb calibration process. Some explanations were made in relation to lower vineyard K_cb values (K_{cb crop}) during the mid-season by Silva et al. [15] and Darouich et al. [16]. The former obtained a good fit parameter, a good r², and a b value close to 1, but crop transpiration was underestimated; conversely, Darouich et al. [16] reported an EF value less than 0.4, especially under ground cover conditions, in Italy. After applying the K_{cb cover adjusted} proposed in the study, an increase in crop transpiration was detected, about 35% on average.

Table 7. Assessment of Equation (9) to adjust K_{cb cover} to real conditions.

Reference	Original			Adjusted
	Kcb crop mid	Kcb cover mid	fc crop mid	Kcb cover mid adj	Kcb crop mid adj
[17]	0.40	0.25	0.53	0.13	0.53
[15]	0.18	0.24	0.40	0.14	0.27
[16]	0.28	0.20	0.29	0.14	0.34
Actual study	0.22	0.24	0.25	0.18	0.27

Note(s): K_{cb crop mid}: basal crop coefficient; K_{cb cover mid}: basal coefficient of active ground cover; f_{c crop mid}: fraction of crop during the mid-season; K_{cb cover mid adj}: basal coefficient of active ground cover using Equation (9); K_{cb crop mid adj}: basal crop coefficient after applying the new approach (Equation (9)). All parameters were referred to mid-season (mid).

summarises the results obtained after the application of the new approach to adjust K_{cb cover} to real conditions (Equation (9)).

In global terms, in all study cases, K_{cb cover} decreased from 30 to 50%, but this is highly dependent on the f_{c crop} value and trellis system (VPS, semi-trellises, etc.). These results support an increase in vineyard transpiration, linked to the real situation of a complex vineyard with active ground cover. The final values (Table 7) agree with the real vineyard water requirements; during the rapid growth stage, vines with a height of 1.0–1.5 m obviously provide higher water requirements than a cover crop with a height of 0.20 m. These limitations of the original methodology [4,11,26] are due to the lower value of K_{cb min} in the K_{cb full} equation. Adaptations to incomplete crops with cover crops have maintained this lower limit, introduced in the original equation for full crops. However, this does not make physical sense in periods where cover crops are decreasing and vineyards are rapidly increasing their vegetative area. Our results show a clear need to reduce the K_{cb cover} values derived with the daily Dual-Kc model using the f_c value of vineyards.

5. Conclusions

Obtaining a good estimate of vineyard water requirements in the presence (continuous or discontinuous) of active ground cover requires thorough adjustment of the parameters affecting the soil water balance. The initial approach defined in the reference studies [4,11,26] resolves the process of determining the transpiration of vineyards and active vegetation (row and/or inter-row) as a whole. However, to differentiate between vineyard transpiration and active ground cover transpiration, a reduction coefficient for the active ground cover transpiration should be adopted to adjust to the ground truth. The initially proposed model limits the minimum value entered for K_{c min} (0.15) to calculate K_{cb cover}, resulting in high values despite the evaporative surface area in the vineyard being larger than that of the vegetation cover. The proposed modification has been tested in previous studies, yielding adequate results for all cases and leading to an average increase of 35% in mid-season vineyard transpiration. The proposed correction factor is easy to apply to row crops with active ground cover, based on the K_{cb cover} value obtained using SIMDualKc and the f_{c crop} value. The presented results demonstrate good applicability to the case under study; as such, the proposed model’s implementation is recommended in similar studies to ensure effective differentiation between crop transpiration and that of active ground cover. However, further research is needed to ensure that the proposed equation is appropriate for different situations around the world.