Low Hanging Fruit’ Practices for Improving Water Productivity of Rainfed Potatoes Under Integration of Cultivar Selection, Mulch Application, and Agroecological Zones in Sub-Tropical, Semi-Arid Regions

1. Introduction

Water productivity (WP) is the ratio of yields or biomass produced per unit of actual water used [1]. Potato (Solanum tuberosum) WP in the sub-Saharan African (SSA) region has been reported particularly in Ethiopia ranging from 3.56 to 8.70 kg m^-3 [2], meanwhile in other regions such as Asia (Pakistan) ranged from 14.60 to 17.28 kg m^-3 [3]. The low potato average yield of 6 to 10 t ha^-1 of SSA [4] is attributed to low WP, exacerbated by high rainfall variability often accompanied by flash floods and intermittent dry spells, which cause water losses and crop water stress during the growing season leading to inefficient water use and decreases in yield [5]. Hence, SSA farming under rainfed is at risk due to climatic variability considering the high dependence on rainfall, also rainfall does not match crop water needs according to each growth stage, often creating a mismatch between the required potato water distribution and occurring rainfall [6]. Potatoes require the least amounts of water from planting to crop establishment (Kc = 0.4 – 0.5) and later towards crop maturity and harvest (Kc = 0.7 – 0.75) [7]. Water requirements gradually increase (Kc = 0.7 – 0.8) during the crop development phase due to vegetative growth and initiation of tuber development, with peak water requirements at the mid-season phase (Kc = 1.05 – 1.2) when peak photosynthesis and tuber bulking occurs [7]. Resource-constrained SA potato producers usually lack irrigation inputs, making potato production under smallholder settings highly reliant on rainfall and exposed to rainfall-related production risks especially when rainfall is a limiting factor [8].

Improving potato WP under rainfed growing conditions is linked to management practices related to changes in crop, soil, and water management. These practices include retaining and storing water during flash flood events, to ensure sufficient water during dry spells. Covering the soil with mulch to conserve soil moisture, increasing infiltration during heavy rain reducing soil evaporation, and restraining runoff [9]. This practice increases in simultaneous increases in crop and water productivity and has been observed to increase WP [10,11,12,13,14] in other settings, however, there is very low use of mulch in SSA by smallholder farmers, particularly in South Africa. Biswal et al. [14] reported that straw mulching significantly increased water productivity (30–39%). Another tool to improve WP is selecting agroecological zones that offer moderate temperatures, adequate rainfall, and fewer extreme events that are accompanied by high atmospheric dryness amplifying dry spell severity and soil moisture deficit during potato production, thereby negatively affecting plant water fluxes and crop yields [15].

Approximately 96% of smallholder farmers use uncertified seed, either potato retained from previous harvest (farmer-saved seed), sourced from neighbors, or purchased from local markets [16]. Uncertified poor-quality seeds from the informal sector, are generally less efficient in converting water into biomass and ultimately into yield. This results in higher usage per unit of yield produced, reducing overall water productivity. To increase the yield of potatoes for each unit of water transpired, the adoption of certified seed from improved cultivars is key to improving WP. Breeders have developed a range of cultivars that match growth cycles with the expected water supply or with the absence of crop hazards, which are tolerant to water stresses such as dry spells [17]. Selecting short to medium-duration cultivars increases WP by escaping late-season water stresses that adversely affect potato growth and development. Meanwhile, late-maturing cultivars can utilize the late-season rainfall. Farmers have a wide range of options in selecting cultivars that are customized to fit local climatic, as well as selecting agroecological that have certain soil types or hydrological properties that can be noted to help a farmer improve WP.

The application of mulch and cultivar selection to improve potato WP in different agroecological zones has been reported by a handful of studies [11,18,19,20]. Despite the solution being put forward such that when using practices solely tends to give small investment returns, for instance, mulch alone does not always guarantee increases because it is scenario-specific (e.g weed control, soil improvement, plant health, moisture retention, etc.) each scenario may require different types of mulch. Cultivar selection alone does not always work out well, due to climate variability, and soils. The sustainable strategy is to combine these practices, such that if one fails one works. This integration of factors has not been studied, especially in the smallholder setting that has low yield translating to low WP. Hence integration approach responds to produce “more crops per drop” such as increasesing output per little water used, recucing losses of water through evaporation and other unproductive process, lastly improve efficiently use of rainfall. Therefore, the specific objectives were to determine (i) understand the seasonal variables influence on seasonal potato water productivity under dryland conditions and (ii) the impact of integration of cultivar selection, application of mulch and locality on potato WP. Further more this study contribute to the selection of cultivars can can improve potato WP and practices that can be used to achieve high yields and WP of potatoes under smallholder settings. In South Africa, there is limited knowledge of the practices that are used by smallholder farmers and quantities of potato yield are very difficult to obtain in this setting, hence this study will cover that research gap.

2. Materials and Methods

2.1. Potato Cultivars Selection

The potato cultivars (Table 1) (Mondial, Sababa, Panamera, and Electra) were sourced from Wesgrow (Pty) Ltd. Cultivars that were selected varied in seasonal length and drought tolerance.

2.2. Field Description and Climate Data

The field trials were conducted in two different agroecological zones in Swayimane (Stezi and Mbhava) and Appelsbosch (Mbalenhle) in KwaZulu-Natal, South Africa. Swayimane and Appelsbosch represented macro-climatically different agroecological. Appelsbosch has cooler, highly humid, drier climates with lower clay content soils that are sandy loam in texture. On the other hand, Swayimane has warmer, moderately humid, wetter climates with high clay content soils that are sandy clay in texture. Swayimane was further divided into two different micro-climatic locations even though they are under the same bioresource groups (BRG), which are Stezi and Mbhava. Stezi is a high-lying (hilltop), warmer, north-facing (sun-facing) slope (Table 2). Ironically, Stezi means high heights which match the description of a high-lying area in this instance. Stezi was expected to have high atmospheric dryness due to its warmer and drier atmosphere compared to Mbhava. Mbhava is a low-lying (valley), cooler, south-facing slope (Table 2). Before initiating planting, automatic weather stations (Davis 6152 Wireless Vantage Pro 2) were installed on the experimental sites, and data such as temperature, rainfall, reference evapotranspiration, and relative humidity were collected. Furthermore, the characterization of soil was done to determine soil forms [23] for each experiment before planting. Afterward SPAW model [24] was used to get the soil hydraulic and physical properties are represented in Table 3.

2.3. Experimental Design and Field Management

Soil samples were collected at each site for soil fertility analysis, analysis was done at Cedara Soil Analytical Services [29] and fertilizer (KCl and 2:3:4 (30)) was applied using on the soil fertility analysis. The field trial was in the 2022/23 and 2023/24 planting seasons (August to January) in the above-mentioned localities (micro-climates). A split-plot layout was used in a randomized complete block design. Grass hay mulch (not mulch and mulch), locality (Mbhava, Stezi, and Mbalenhle), and cultivars (Panamera, Electra, Sababa and Mondial) were tested. The pre-sprouted seeds were manually planted on a 13.5 m² plot with a spacing of 0.3 m x 0.9 m. Upon crop establishment stage 1.9 kg m^-2 mulch was applied evenly. Dithane M45 (3 kg ha^-1), Ridomil (2.5 L ha^-1) and Bravo (1 L ha^-1), Bravo (1 L. ha^-1) fungicides were used weekly to control blight diseases upon disease incident, insect was controlled as well using Cypermethrin and Decis at a rate of 0.150 L. ha^-1 [30]. Weed was manually removed during the growing season, and round-up (5 L. ha^-1) was only used before plowing the soil [30].

2.4. Collection of Field-Measured Variables

2.4.1. Growth and Phenological Parameters

The canopy cover was collected using the Canopeo v2.0 application, the captured high-resolution image, the application evaluated pixels of the image based on the red-to-green (R/G), and blue-to-green (B/G) color ratios [31]. The result output image is in black and white, the green area is represented by white, and black is the non-green area. Thereby maximum canopy cover was the highest percentage of canopy cover, meanwhile duration of maximum canopy cover was the number of days between maximum canopy decline and days when maximum canopy cover was reached. Phenological data included time taken to reach physiological maturity and was measured when 50% of plants per plot showed some signs of yellowish leaves (senescence). At harvest, yield was measured on the experimental plot, two-center rows.

2.4.2. Crop Water Use

Weekly soil water content (SWC) was measured using a Diviner 2000 (Sentek Environmental Technologies, Stepney, Australia) through PVC access tubes that were installed on the day of planting at the center of the plot, and the first reading was taken on the same day. The reading was done on different depth intervals of 10, 20, 30, 30, 40, 60, 70, 80, 90, and 100 cm as the percentage of soil water content. Furthermore, measured SWC was used to calculate soil water balance from planting to physiological maturity as shown below.

ET = P - ∆SWC

(1)

where ET is actual crop water use, P is rainfall and SWC is the change in soil water content (soil water content at harvest - soil water content at planting). The study was conducted under rainfed conditions hence irrigation was ignored, and runoff, capillary rise, and drainage were negligible. The cultivars reached physiological maturity on different days, hence rainfall used to calculate ET was based on days to reach maturity. Thereby water productivity (WP) was calculated using the following equation:

WP= Yield/ actual water use

(2)

2.4.3. Vapor Pressure Deficit

The vapor pressure deficit was calculated with the following equations (3) to (6) [32]:

$$\mathrm{e}°\left(\mathrm{T}\right)=0.6108\mathrm{e}\mathrm{x}\mathrm{p}\left[{\displaystyle \frac{17.27\mathrm{T}}{\mathrm{T}+273.3}}\right]$$

(3)

$${\mathrm{e}}_{\mathrm{s}}={\displaystyle \frac{{\mathrm{e}}^{°}\left({\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\right)+\mathrm{e}°\left({\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)}{2}}$$

(4)

$${\mathrm{e}}_{\mathrm{a}}={\displaystyle \frac{\mathrm{e}°\left({\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)\frac{{\mathrm{R}\mathrm{H}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}{100}+\mathrm{e}°\left({\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\right)\frac{{\mathrm{R}\mathrm{H}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}{100}}{2}}$$

(5)

$$\mathrm{V}\mathrm{P}\mathrm{D}={\mathrm{e}}_{\mathrm{s}}-{\mathrm{e}}_{\mathrm{a}}$$

(6)

where e° (T) is the saturation vapor pressure at air temperature T (°C), Tmax is the daily maximum temperature, Tmin is the minimum temperature, RHmax is the maximum relative humidity, RHmin is the minimum relative humidity, es is saturated vapor pressure, ea is actual vapor pressure and VPD is vapor pressure deficit.

2.4.4. Statistic Analysis

GenStat software (GenStat®, 23^rd edition [64 Bit] VSN International, UK) was used to analyze data. Tukey was used for the mean separation test at p<0.05. The association of seasonal variables was correlated using GenStat.