Seasonal Variations in Water Use of Japanese Plum Orchards Under Micro-Sprinkler and Drip Irrigation Methods Using FruitLook Data

Abstract

South Africa is considered one of the driest countries, and its water insecurity challenges are exacerbated by climate change and variability, depletion, and degradation, among other factors. The challenges of water insecurity are exacerbated by some of the introduced crops, like the Japanese plums (Prunus salicina Lindl.) grown in South Africa, as they consume a lot of water. The Japanese plums are grown under irrigation to supplement low and erratic rainfall in the country. There is little information on the water requirements of Japanese plums (particularly in water-scarce regions), a gap addressed by this study. Therefore, the study aims to quantify and compare the seasonal water use of high-performing, full-bearing Japanese plum orchards under drip and micro-sprinkler irrigation in the Western Cape Province, using readily available satellite data from the FruitLook platform. The seasonal water use volumes of selected plum orchards were compared at provincial and farm scales. At a provincial scale, micro-sprinkler-irrigated orchards consumed significantly more water (up to 19%) than drip-irrigated orchards, whilst drip-irrigated orchards experienced an average 38% greater water deficit. Results were more variable at the farm scale, which was attributed to the influence of site-specific soil, climate, and crop conditions on the performance of the irrigation methods. Therefore, a blanket approach cannot be used when selecting an irrigation method and design. Instead, a case-by-case approach is recommended, which takes into account the root distribution, soil texture, and planting density, among other factors. The generated knowledge facilitates allocating and licensing water resources, developing accurate irrigation scheduling, and promoting improved water use efficiency.

1. Introduction

Effective water management in modern agriculture is critical in water-scarce regions where the sector faces intense competition for water. The South African deciduous fruit industry, as with other key fruit-producing countries in Mediterranean climates, is heavily reliant on irrigation to supplement the country’s seasonal and erratic rainfall [1]. Irrigated agriculture remains the largest water user in South Africa (accounting for over 60% of total consumption), with demand for agricultural produce projected to increase in the coming years due to population growth [2]. There is, therefore, an increasing need to improve water use efficiency (WUE) in the agriculture sector. The challenges of water scarcity are further exacerbated by climate change and increasing inter- and intra-sectoral competition. This can be achieved by either increasing yield without increasing water use or reducing water use without decreasing yield and fruit quality [3]. However, in water-scarce countries like South Africa, where water availability is one of the primary limiting factors for extensive agricultural expansion and sector sustainability, the latter approach is more ideal [4].

The sustainability of the agricultural sector, through optimizing crop water while improving crop yield, can be achieved by adopting efficient irrigation systems. The selection of the appropriate irrigation method significantly influences water consumption and plant health and quality [5]. In recent years, precision irrigation technologies have been widely adopted (i.e., micro-sprinkler and drip methods) in South African orchards with the aim of reducing water use and input costs whilst improving WUE and economic returns [6]. The water-saving potential of drip and micro-sprinkler irrigation methods compared to conventional flood and surface methods has been well-documented in the literature [7,8,9,10,11,12]. However, Ntshidi et al. [13] highlighted the importance of adequate system design and irrigation scheduling to achieve optimal results. Micro-sprinkler and drip methods are precision irrigation systems that facilitate localized irrigation and fertigation. However, they primarily differ in terms of the emitter design, the size of the wetted area, and the application efficiency (

Table A3. Principal differences, advantages, and disadvantages of drip and micro-sprinkler irrigation methods.

	Drip	Micro-Sprinkler
Water distribution	Water is delivered directly to each plant’s rootzone, slowly and precisely	Water is distributed above the soil surface, over a wide area in a circular pattern
Water application rate	Lower application rate	Higher application rate
Area coverage	Localized coverage while targeting the root zone	Broader area while wetting the soil surface around each plant
Soil moisture distribution	Localized and a deeper wetting pattern	Broader and shallow wetting pattern
System design	Requires more complex design and installation (precise emitter positioning and flow rate calibration)	Easy installation and maintenance (few components and less precise placement needed)
Advantages	Minimized water loss through runoff, evaporation, and deep percolation	Uniform soil surface coverage, which is beneficial for extensive root systems
	Weed growth is reduced in the area between plants	Reduces plant stress by cooling the soil and plants in hot weather
	It lowers the risk of moisture-related plant diseases, such as fungal diseases, by keeping the foliage dry	Protects plants from frost by creating a thin layer of water on the plants, which releases latent heat upon freezing
	Fertilizer is applied directly and precisely where needed	Less clogging due to larger water passage
Disadvantages	Installation costs due to specialized tubing, emitters, and filters	Increased wind drift and evaporation
	Emitters clog due to inadequate water filtration	Increased risk of fungal diseases from watering the whole plant
	Complex system design that requires precise planning and management	Weed growth around the plants due to broader water distribution

).

In drip irrigation systems, water is applied (typically at low volumes and high frequency) directly to the tree root zone through drip emitters spaced at predetermined distances. This minimizes non-beneficial orchard water use (evaporation from soil and transpiration from cover crop), thus promoting greater water application efficiency [14]. Contrarily, micro-sprinkler systems apply water via sprinkler jets (higher volumes at a reduced frequency) over a greater radius, producing a larger wetted surface area. Consequently, the larger wetted area reduces the application efficiency of the method as a more significant portion of water is lost through non-beneficial water uses [1,15]. The differences in the principles of operation of these two irrigation methods and their advantages and disadvantages are shown in Table A3 (Appendix A). Differences in the wetting pattern of a soil profile under drip and micro-sprinkler irrigation are prominent. Irrigation in drip systems is more localized, promoting a narrow and deep wetting pattern, whereas the larger application radius in micro-sprinkler systems promotes a laterally wider but shallow pattern. Soil texture, organic matter, and stone content also greatly influence the movement of water through the soil profile, with a narrower, deeper shape expected in sandy soil with high stone content (low water retention capacity) compared to a shallow, wider shape in clay-rich soils with higher organic matter content (high water retention capacity) [16,17].

Comparative analyses of the influence of micro-sprinkler and drip irrigation methods on orchard water use and water use efficiency have been conducted for various key fruit crops, namely apples [18,19,20] citrus [12], peaches [9], pears [21], cherries [22,23], avocadoes [24], mangoes [25,26], and bananas [27,28] Most of these studies reported that tree water deficit, particularly during the crucial phenological growth stages, has been shown to negatively impact yield and fruit quality. According to the FAO Irrigation and Drainage Paper No. 66 [29], the most sensitive stages to water deficit are flowering, fruit set, and fruit enlargement. During the flowering and fruit set stages, water deficits may lead to reduced pollination and fruit set while lowering initial fruit growth, directly affecting yield potential. Water shortages during the fruit enlargement stage (usually coinciding with the highest ET demand) impact fruit size (small fruit) and quality (physiological disorders). Contrariwise, pre- and post-harvest periods are generally less sensitive to water deficits, although they may affect bud development and vegetative growth for the next season. Therefore, it is imperative to tailor irrigation management during these critical phenological stages to ensure high fruit quality and sustainable production [30]. Prioritizing water supply during these sensitive stages aligns with the recommendations of the FAO Irrigation and Drainage Paper Number 66 [29].

Ntshidi et al. [13] reported reduced water use in a single-line drip-irrigated apple orchard. However, the trees experienced a more significant water deficit than those under micro-sprinkler irrigation, which resulted in smaller canopies, reduced stomatal conductance and transpiration rate, and inferior fruit quality and yield. Teixeira et al. [10] observed similar results where a drip-irrigated lemon orchard had higher water use efficiency (reduced consumption) than a micro-sprinkler-irrigated orchard but exhibited more significant water stress. Lebese et al. [19] and Li et al. [26] noted contrasting findings in apple and young mango orchards, respectively, where drip-irrigated trees exhibited higher photosynthetic rates and stomatal conductance along with lower water use than micro-sprinkler-irrigated orchards. This resulted in higher yields, fruit quality, root development, and water use efficiency in drip-irrigated orchards. The opposing findings in the literature and the lack of quantitative comparative analyses create uncertainty around the best irrigation method, resulting in farmers haphazardly switching from one irrigation method to the other on a trial-and-error basis.

In recent years, remote sensing techniques have been shown to have the ability to estimate crop water use to manage irrigation at different temporal and large spatial scales. In this study’s companion paper by Mashabatu et al. [30], FruitLook (an online platform with underpinning models, ETLook and SEBAL) was used to estimate water use of a full-bearing Japanese plum orchard in Robertson, South Africa. The selected orchard in the study was assumed to be optimally irrigated. FruitLook’s estimates (seasonal = 744 mm, annual = 948 mm) were validated using the micrometeorological eddy covariance (EC) system’s measurements (seasonal = 751 mm, annual = 996 mm). Although FruitLook [31] correlated well with the EC system with a Nash–Sutcliffe Efficiency (NSE) of 0.91, it slightly underestimated plum water use by a Pbias of 6.15%. The EC system’s and FruitLook’s measured and estimated volumes were within the range of Japanese plums’ water use volumes reported by Mashabatu et al. [32]. Mashabatu et al. [30] therefore demonstrated the potential of FruitLook to provide reliable means to estimate water use in full-bearing and high-density plum orchards.

This study compared the seasonal water use (ET) and water deficit of full-bearing Japanese plum orchards under drip and micro-sprinkler irrigation in two major production regions (Robertson and Wellington) in the Western Cape Province of South Africa using readily available satellite data from the FruitLook platform (https://FruitLook.co.za/, accessed on 1 October 2024). Besides providing crop–water productivity information on the two irrigation types, this study aims to provide invaluable outcomes to guide farmers in selecting the correct irrigation method and design to improve orchard water use efficiency without negatively impacting fruit yield and quality. This study’s limitation is the lack of access to harvest data, which inhibits direct calculation of water productivity or crop–water productivity as the yield produced per unit water applied or consumed.

This study hypothesizes that the differences in water use in the selected orchards (commercial, well-managed, and optimally irrigated) are primarily attributed to the chosen irrigation system. This assumption can also be preserved as a second limitation of the study. However, the main purpose of the study was to make use of the satellite resource to make a comparison of consumptive water use of plums under drip and micro-jet irrigation. It is acknowledged that each orchard has its own specific environmental and management conditions, but the point was not to descend into that level of detail. The assumption was that the large volume of information obtainable from satellite monitoring would average out specific orchard conditions.

2. Materials and Methods

2.1. Selection of the Study Orchards

Georeferenced spatial data depicting the distribution of Japanese plum orchards across the Western Cape Province along with the employed irrigation method were extracted from the CapeFarmMapper platform (https://gis.elsenburg.com/apps/cfm/, accessed on 1 June 2022). Most of the orchards were situated within the agriculturally prominent Cape Winelands district, with the rest falling within the Central Karoo, City of Cape Town, Garden Route, Overberg, and West Coast districts. According to the metadata, the field boundaries were digitized (mapped) during the 2017/18 agricultural season using aerial photographs from 2016. There was no indication that the dataset had been updated since then; therefore, it was assumed that some of the delineated orchards were likely outdated. This necessitated the ground-truthing of a narrowed pool of orchards selected for validation and use in this study. Farmers’ participation was very important, as we required access to the farms to verify the crop type and the employed irrigation method. Several farm managers declined to participate in the study or did not respond to calls and emails, reducing the pool of selectable orchards. Additionally, orchards under agricultural nets, i.e., not under natural open field conditions, were omitted.

A ground-truthing campaign was carried out in the Wellington and Robertson regions from August to November 2022, highlighting multiple errors in the extracted dataset. Firstly, abandoned plots, fallow land, and other crop types (i.e., citrus, nectarines, peaches, etc.) were demarcated as plum orchards. In other instances, the spatial dataset did not include plum orchards found on the ground. Additionally, the listed irrigation method at multiple orchards, particularly in Wellington, was incorrect, i.e., micro-sprinkler-irrigated orchards were listed as drip-irrigated and vice versa. Lastly, the field boundaries of some orchards were not properly digitized (i.e., the digitized plots included bare soil and other land surfaces surrounding the orchards). A total of 150 plum orchards, cultivated on 239.5 hectares of land and spread across 11 farms, were validated at the end of the ground-truthing campaign. Twenty-four cultivars were planted in these orchards, with Angeleno, Ruby Sun, Sunkiss, and Fortune cultivars being the most prominent. The trees were trained on a single- or dual-row palmette trellis system. The age of an orchard was used to determine its bearing status (bearing or non-bearing). Plum trees typically begin bearing fruit from 3 years. Therefore, orchards older than 3 years were considered full-bearing. In instances where orchard age was not readily available, visual observations of tree features (tree height, stem thickness, canopy density, and the presence of fruits) and communication with farm workers in comparison with features of known bearing trees were used to determine the bearing status of observed trees. Examples of trees used as benchmarks for comparison and selected orchards using this method are shown in Figure 1. A total of 49 orchards were selected using this method, 14 in Robertson and 35 in Wellington. A second shapefile was created, where the validated plum orchards were digitized, and the employed irrigation method was noted. Additional information on the selected orchards (cultivar, age, size, and training system) is presented in Appendix A.

The selected orchards in Robertson were situated on six farms along the valley floor. These were the Smuts Brothers (Klipboschlaagte and Lucerne), Mon Don, Sonskyn, Rosedale, and Ebendale farms (Figure 2). There were 76 orchards planted on 126.8 hectares of land in total, giving an average area of 1.67 ha (

Table 1. Summary of farms and plum orchards selected in Robertson and Wellington (South Africa) for the comparison between micro and drip irrigation water use.

Region	Farm	No. of Orchards	Area Under Cultivation (ha)	Irrigation Type
Robertson	Ebendale	18	23.3	Drip
	Klipboschlaage	9	30.1	Drip
	Lucerne	6	12.4	Drip
	Mon Don	20	28.4	Drip and micro-sprinkler
	Rosedale	15	16.9	Drip
	Sonskyn	8	15.7	Drip
Wellington	Abendruhe	8	19.1	Micro-sprinkler and drip
	De Goede Hoop	15	10.1	Micro-sprinkler
	Louisvale	12	12.2	Micro-sprinkler
	Sandrivier	32	57.2	Micro-sprinkler
	Welgemoed	7	14.1	Micro-sprinkler
	Total	150	239.5

). The farms, except for Ebendale, were within 5 km of each other. Ebendale is in Bonnievale, a neighboring town approximately 20 km from Robertson. Most of the farms in the area use drip irrigation, with Mon Don being the only exception, with 13 out of the 20 orchards being irrigated via micro-sprinklers. In terms of area under cultivation, the Klipboschlaagte farm was the largest (30.1 ha), followed by Mon Don (28.4 ha) and Ebendale (23.3 ha).

In Wellington, the selected orchards were situated on 5 farms within a 4 km radius of each other. These were the Sandrivier, De Geode Hoop, Louisvale, Abendruhe, and Welgemoed farms (Figure 3). In total, 74 selected orchards were cultivated on 112.7 hectares of land (Table 1). Orchards in the region were predominantly irrigated via micro-jets, with the Abendruhe farm being the only farm where both irrigation methods (drip and micro) were employed.

2.2. FruitLook

FruitLook data (https://FruitLook.co.za/, accessed on 1 October 2024) were used to estimate and compare water use volumes of micro-sprinkler and irrigated Japanese plum orchards in Wellington and Robertson. Water use (ET) volumes of these orchards are computed using the ETlook model [31], satellite imagery from Sentinel, Landsat, VIIRS, and MODIS, and weather data from ground-based automatic weather stations (AWSs) at a 10 m spatial resolution. As described in the companion paper by Mashabatu et al. [30], ETLook estimates of evaporation (E) and transpiration (T) are estimated separately using the two-layer Penman–Monteith equation [33] and integrated transport resistances. The equation is written as follows:

$$E={\displaystyle \frac{∆\left(R_{n,soil}-G\right)+\rho C_p\left(\frac{{∆}_e}{r_{a,soil}}\right)}{∆+\gamma \left(1+\frac{r_{soil}}{r_{a,soil}}\right)}}$$

(1)

$$T={\displaystyle \frac{∆\left(R_{n,canopy}\right)+\rho C_p\left(\frac{{∆}_e}{r_{a,canopy}}\right)}{∆+\gamma \left(1+\frac{r_{canopy}}{r_{a,canopy}}\right)}}$$

(2)

where Δ is the slope of the saturation vapor pressure curve; Δ_e is the vapor pressure deficit; $\rho$ is the air density; C_p is the specific heat of dry air; γ is the psychometric constant; R_n,soil and R_n,canopy are net radiations of the soil and the canopy, respectively; r_soil and r_canopy are resistances of the soil and the canopy, respectively; and r_a,soil and r_a,canopy are aerodynamic resistances of the soil and the canopy, respectively. The exact procedure (flowchart) used by FruitLook to calculate ET is not readily available to the public as it is protected by Intellectual Property (the underpinning model is licensed by eLeaf, the owner of SEBAL). FruitLook calculates the ET deficit as the difference between the optimal reference evapotranspiration (ETo) and the actual ET rate for a given orchard. ETo is calculated using the Penman–Monteith equation [33]. FruitLook’s temporal coverage varies from season to season. Water use estimates and water deficit data were extracted from the Fruitlook portal for the 2018, 2019, 2020, 2021, and 2022 seasons.

2.3. Statistics

An independent Student’s t-test, at a 5% confidence level (α = 0.05), was performed to test if there was a statistically significant difference between the estimated water use (ET) of micro-sprinkler- and drip-irrigated Japanese plum orchards. The null (Ho) and alternate (Ha) hypotheses were the following.

Ho (p > 0.05): There is no statistically significant difference between the estimated water use in micro-sprinkler- and drip-irrigated Japanese plum orchards.

Ha (p < 0.05): There is a statistically significant difference between the estimated water use in micro-sprinkler- and drip-irrigated Japanese plum orchards.

Student’s t-test has four primary assumptions, as follows: (1) the data are continuous, (2) the data were sampled from a random population, (3) the data are normally distributed (follow a bell-shaped curve), and (4) the sampled datasets have equal or similar variances. While the first 2 assumptions were satisfied, a normality test (Shapiro–Wilk) and f-test were conducted to test the validity of the third and fourth assumptions, respectively.

3. Results

3.1. Water Use of Micro-Sprinkler- and Drip-Irrigated Plum Orchards (2022/23 Season)

FruitLook-derived estimates were used to compare the water use and water deficit of full-bearing Japanese plum orchards under micro-sprinkler (Wellington) and drip (Robertson) irrigation methods from 1 August 2022 to 31 July 2023 (regarded as the main season, as it had the largest selection of orchards under consideration). Weekly estimates were aggregated to an annual time step and compared at a regional level. In this comparison, micro-sprinkler-irrigated orchards at Mon Don farm in Robertson (13 orchards) and drip-irrigated orchards at Abendruhe farm (3 orchards) in Wellington were omitted. The large-scale comparison looked at the differences in water use of drip-irrigated orchards in Robertson (hence only the drip-irrigated orchards at Mon Don were included, and micro-sprinkler-irrigated orchards were omitted) and micro-sprinkler-irrigated orchards in Wellington (micro-sprinkler orchards were included whilst drip orchards were omitted). Figure 4 shows a histogram of the water use (ET) estimates, depicting a normal distribution, with a high frequency of the observations occurring between 1000 and 1099 mm for micro-sprinkler-irrigated orchards (44% of observations) and 950 and 1049 mm for drip-irrigated orchards (56% of observations). Micro-sprinkler-irrigated orchards consumed significantly more water than drip-irrigated orchards (p < 0.01), with mean water use estimates of 1019 and 961 mm a⁻¹, respectively (a 6% difference). The mean and median water use estimates for orchards under both irrigation methods were similar (<2% difference).

Figure 5 shows a histogram of water deficit estimates. Mean water deficit estimates were relatively low, ranging from 0 to 32 mm in micro-sprinkler-irrigated orchards and between 2 and 47 mm in drip-irrigated orchards. The highest frequency of observations was between 0 and 13 mm (96% and 56% of observations for micro-sprinkler- and drip-irrigated orchards). Thereafter, a decreasing trend with an increasing value increment was observed. The relatively low ET deficit estimates suggest that most orchards experienced minimal water stress and, as such, were optimally irrigated. However, deficit estimates were substantially higher in drip-irrigated orchards (by 178%), indicating greater water stress compared to micro-sprinkler-irrigated orchards.

The same comparison was conducted at the farm scale using estimates from the Mon Don (Robertson) and Abendruhe (Wellington) farms (

Table 2. Water use (ET) and water deficit estimates for micro-sprinkle- and drip-irrigated orchards at the Mon Don (Robertson) and Abendruhe (Wellington) farms.

	Mon Don		Abendruhe
Parameters	Micro-Sprinkler	Drip	Micro-Sprinkler	Drip
No. of orchards	13	7	5	3
ET mean (mm a−1)	999	1001	917	877
ET median (mm a−1)	990	982	908	879
Water deficit (mm a−1)	5	14	10	5
StDev	60	71	83	44
CV (%)	6	7	9	5

Note: StDev—standard deviation; CV—coefficient of variation.

). Orchards at these farms were irrigated using both micro-sprinkler and drip irrigation methods. It was understood that farm-scale results, due to a smaller sample size, are likely to be less representative of regional water use dynamics, as the influences of orchard management practices and site-specific conditions are expected to be more prevalent. Nonetheless, these estimates could provide valuable insights into orchard water use dynamics under similar growing conditions. Akin to the regional comparison, micro-sprinkler-irrigated orchards consumed more water than their drip-irrigated counterparts, with 0.7% and 3% higher consumption at the Mon Don and Abendruhe farms, respectively. The water deficit estimates at Mon Don were comparable to regional estimates (30% higher under drip irrigation), although the order of magnitude of the difference was smaller. The opposite was observed at Abendruhe, where micro-sprinkler-irrigated orchards experienced greater water stress than those under drip irrigation. Water deficit estimates were highly variable, particularly at the regional scale, where the coefficient of variation (CV) was 78 and 216% for drip- and micro-sprinkler-irrigated orchards, respectively. Lower CV values (<60%) were observed at the farm scale.

3.2. Seasonal Water Use of Drip- and Micro-Sprinkler-Irrigated Plum Orchards (2018/19–2022/23 Season)

The annual water use and water deficit of micro-sprinkler- and drip-irrigated orchards from the 2018/19 to the 2022/23 season (five seasons) were compared using the same methodological procedure described previously in Section 3.1. Orchards younger than 3 years (deemed not to be full-bearing) and orchards where the planting year was not available (orchards selected based on visual observation) were omitted in seasons preceding the 2022/23 season (2018/19 to 2021/22). The latter were omitted because without sound knowledge of the age of the orchard, it would not be possible to determine the bearing status of orchards in previous seasons accurately. Young, non-bearing trees have been recorded to use significantly less water than mature, full-bearing trees [6,15], and, therefore, their inclusion would add an element of uncertainty to the estimated water use values.

Table 3. Number of orchards under micro-sprinkler and drip irrigation from the 2018/19 to 2022/23 season.

Season	Micro-Sprinkler	Drip	Total
2022/23	71	63	135
2021/22	47	35	82
2020/21	29	41	70
2019/20	24	40	64
2018/19	19	31	50

presents the number of orchards available for comparison in the respective seasons.

Seasonal mean ET and water deficit estimates are presented in

Table 4. Seasonal mean water use (ET) estimates of micro-sprinkler- and drip-irrigated orchards at the regional scale. The highlighted row (2021) was excluded from long-term (2018–2022) average calculations.

Season	Micro	Drip	Difference (%)
2018	1019	948	7%
2019	1077	886	19%
2020	944	810	15%
2021	782	791	1%
2022	1019	961	6%
Average	968	879	10%
Stdev	114	78
CV	12%	9%

Note: StDev—standard deviation; CV—coefficient of variation.

and Figure 6. There was minimal variation in annual water use estimates (2018 to 2022), with a CV value of less than 10% for orchards under both irrigation methods. Estimates varied between 879 (2021) and 1086 (2019) mm a⁻¹ for micro-sprinkler-irrigated orchards and between 797 (2021) and 974 (2022) mm a⁻¹ for drip-irrigated orchards. The low water use estimate for drip-irrigated orchards in 2021 was due to 2 months of missing data (August and September) and, as such, was excluded when calculating long-term averages. Micro-sprinkler-irrigated orchards consistently consumed more water than drip-irrigated orchards, with long-term water use figures of 1003 and 914 mm a⁻¹, respectively (9% difference). The largest differences in water use occurred during the 2019/20 and 2020/21 seasons, with deviations of 19% and 15% (p < 0.01). Conversely, higher water deficit estimates were observed in drip-irrigated orchards (on average 38% higher), with the 2021/22 season being the only exception, potentially due to the period of missing data at the beginning of the season. There was a negative correlation between the percentage differences and the obtained p-values, where a larger difference correlated to a smaller p-value (Figure A1). This is because larger differences are more likely to be statistically significant than smaller differences.

Contrasting findings were observed at the Mon Don (Robertson) farm, as drip-irrigated orchards consumed, on average, 3% more water than micro-sprinkler-irrigated orchards, with a peak deviation of 11% occurring during the 2020/21 season (

Table 5. Summary of seasonal mean water use (ET) estimates under micro-sprinkler and drip irrigation at Mon Don farm (Robertson).

Season	Micro	Drip	Difference
2018	1073	1104	3%
2019	928	988	6%
2020	840	799	5%
2021	813	800	2%
2022	999	1001	0.1%
Average	931	938	1%
Stdev	108	135
CV	12%	14%

Note: StDev—standard deviation; CV—coefficient of variation.

). Average median water use estimates from 2018/19 to the 2022/23 season for micro-sprinkler- and drip-irrigated orchards were 928 and 959 mm a⁻¹, with a CV of 12% and 11%, respectively. Water use of drip-irrigated orchards was higher at Mon Don (959 mm a⁻¹) compared to regional estimates (914 mm a⁻¹). This is likely a result of the larger sample size at the regional scale, which encompasses orchards under variable growing conditions and thus a wider range of water use estimates, which are more so on the lower end of the spectrum. Additionally, water deficit estimates were marginally higher in drip-irrigated orchards, with an average difference of 1% (Figure 7). An inverse relationship between water use and water deficit can be seen, where maximum water deficit estimates coincide with minimum water use estimates in both micro-sprinkler- and drip-irrigated orchards. This is most evident during the 2021/22 and 2022/23 seasons at both the regional and the farm (Mon Don) scale, where the maximum water deficit in 2021/22 coincided with reduced water use, followed by a sharp decline in the following season (2022/23) coupled with an increase in water use. Six of the eight orchards at the Abendruhe farm were omitted in the seasons preceding the 2022/23 season according to the omission criterion. Therefore, a long-term, farm-scale water use comparison in Wellington was not possible.

3.3. Statistical Analysis

The results of the Shapiro–Wilk, f, and t tests are presented in

Table 6. Statistical analysis results (p-values) for the normality, f, and t tests, where the p values highlighted in orange are > 0.05 (Ho) and the values in red are < 0.05 (Ha).

Season	Shapiro–Wilk Test		f-Test	t-Test
	Micro-Irrigation	Drip Irrigation
2018	0.06	0.74	0.77	0.03
2019	0.07	0.43	0.79	3.23 × 10−12
2020	0.24	0.18	0.96	1.02 × 10−5
2021	0.66	0.21	0.02	0.44
2022	0.08	0.06	0.06	7.63 × 10−4

. In the case of the Shapiro–Wilk test, all p-values were above the significance threshold, which suggests that the datasets are normally distributed. The f-test p-values follow a similar trend (p > 0.05; equal variances between datasets), except for the 2021 result, where p = 0.02. In this case, an unequal variance t-test was conducted. All but one (2021) of the t-test p-values were below the significance threshold (p < 0.05). This infers that there is a statistically significant difference between the estimated water use of selected micro-sprinkler- and drip-irrigated plum orchards during the 2018, 2019, 2020, and 2022 seasons. The p-value for the 2021 season (p = 0.44) exceeded the significance threshold, suggesting that the observed differences in water use were not statistically significant. The deviation of the t-test p-values from the significance threshold is shown in Figure 8.

4. Discussion

This study produced contrasting findings on water use and water deficit which varied at regional and farm scales over the study period. At a regional scale, micro-sprinkler-irrigated orchards consumed significantly more water (up to 19%), whilst ET deficit estimates were 38% higher in drip-irrigated orchards. At the Mon Don farm, water use was higher in drip-irrigated orchards, whilst the difference in ET deficit estimates was marginal (1% difference). Conversely, micro-sprinkler-irrigated orchards at the Abendruhe farm exhibited a greater water deficit than drip-irrigated orchards despite having higher water consumption. Results at regional scale are in line with findings by Ntshidi et al. [6], [15] and Teixeira et al. [10], where drip-irrigated apple and lemon orchards used less water, but experienced greater water deficit stress compared to micro-sprinkler-irrigated orchards. Given the larger sample size (n = 135 in the 2022/23 season), the regional scale comparison provided a more representative depiction of orchard water use dynamics under both irrigation methods in each area.

However, contradictory results on the farm scale suggest that site-specific conditions largely impact the performance of drip and micro-sprinkler irrigation methods at the orchard scale. These include irrigation system design, irrigation scheduling, orchard management practices, soil texture, etc. The impact of a chosen irrigation method (drip or micro-sprinkler) on tree water status and, subsequently, water deficit is likely to differ from one orchard to another (even on the same farm) due to the influence of these factors. This assumption is corroborated by contrasting results observed in the literature. For example, Lebese et al. [19], Fallahi et al. [20], and Li et al. [26] reported increased yield and fruit quality in drip-irrigated fruit orchards, with no indication of significant water deficit stress, in contrast to findings by Teixeira et al. [10] and Ntshidi et al. [13]. A commonality in these studies is that the tree water status was largely affected by water availability in the soil profile for root uptake. Water movement through the soil profile differs under drip and micro-sprinkler irrigation, which ultimately affects the root distribution and soil water availability [34]. Irrigation in drip systems is more localized, promoting a smaller, narrower, and deeper wetting pattern, whereas the larger application radius in micro-sprinkler systems promotes a laterally wider pattern. Vercrumbre et al. [35] modeled the root distribution of a plum rootstock grafted to a peach scion in silty clay loam soil. They found that the plum root system exhibited a shallow and horizontal growth pattern from the tree trunk. Ntshidi et al. [13] noted a similar feature in apple orchards in the Western Cape. Water deficit occurs when the wetted soil area does not sufficiently enclose the root system to meet the plant’s water demand. Therefore, it can be argued that the wetting pattern under micro-sprinkler irrigation facilitates greater water availability for root uptake, thus promoting better tree water status.

Using a modified soil–plant–atmosphere continuum (SPAC) model, Garcia-Tejera et al. [36] assessed the influence of the wetted area size on the transpiration rate of a drip-irrigated olive orchard. Despite optimal irrigation scheduling, they concluded that the smaller wetted area under drip irrigation limited maximum tree transpiration. Espadafor et al. [37] and Roble et al. [38] reported similar findings, where an increase in the wetted area culminated in increased transpiration rates and improved tree water status compared to trees with a smaller wetted area. While increasing the size of the wetted area by either converting from drip to micro-sprinkler methods [13,33] or adding more driplines and emitters per tree [38] improved tree water status, orchard yield, and fruit quality, it should also be noted that a larger wetted area is associated with increased orchard floor evaporation [6,12,15,25,26]. Therefore, designing and implementing precision irrigation systems requires a detailed understanding of trees’ physiological responses to irrigation to minimize water consumption while maximizing productivity.

5. Conclusions

This study has provided a comparison between the water-saving potential of drip irrigation and micro-sprinkler irrigation. However, the potential limitations of drip systems were highlighted, and emphasis was put on the need for adequate design and implementation of precision irrigation technologies to maximize water use efficiency without negatively impacting yield and fruit quality. Additionally, it was noted that orchard responses to a specific irrigation method were inconclusive and variable at the farm scale, indicating the influence of site-specific conditions on irrigation system performance. Therefore, a blanket approach cannot be used when selecting an irrigation method and design. Instead, a case-by-case approach is advised, which takes into account the root distribution, soil texture, and planting density, among other factors.

Findings from this study showed that FruitLook has great potential as a monitoring tool to (1) identify weak spots within an orchard, (2) measure water use to meet usage targets, (3) evaluate historic water use (previous seasons), and (4) compare orchards, including different orchards within the farm. Overall, FruitLook can be used in conjunction with traditional techniques to improve orchard and water use management, which can ultimately increase profit margins (through reduced water use, e.g., lower water tariffs, and increased yield, e.g., identification and remediation of weak spots).