Extending the Capsicum Growing Season under Semi-Arid Climate by Using a Suitable Protected Cropping Structure ^†

Abstract

Carnarvon is a key horticultural district in Western Australia which is located approximately 900 km north of Perth and is characterised by a semi-arid climate. In Carnarvon, capsicum (Capsicum annuum L.) is the second most important vegetable crop after tomato, with approximately 3700 tonnes of capsicum fruit produced annually with a farm gate value of AUD 13.5 million. High temperatures, excessive sunlight, low air humidity, and strong wind in spring and summer are major impediments to the achievement of high yield and quality of capsicum in this region. Capsicums are usually planted between March (early autumn) and May (late autumn), and the harvest is usually finished by October (spring) of the same year when grown under shade net houses. However, the internal microenvironment in the shade net houses is sub-optimal for the crop in the early and late growing season due to excessive temperatures and low humidity, resulting in a shorter harvest window and lower production. This study was conducted to examine the possibilities to extend the cropping season for capsicum varieties (i.e., Chevello and Chevi) grown under the retractable roof production system (RRPS) and explore an alternative protected cropping structure that is more affordable and suitable to grow vegetable crops under Carnarvon weather conditions. Overall, the results showed that capsicums planted in February (planting 1) performed better than specimens planted later on in the season: planting 1 performed better and yielded the highest marketable fruit yield (102.6 t ha⁻¹) compared to those planted in early April (planting 2, 72.5 t ha⁻¹) and late May (planting 3, 36.1 t ha⁻¹). The RRPS effectively mitigated the adverse weather conditions and provided a more optimised internal microenvironment for vigorous crop establishment in late summer and an extended harvest in late spring, leading to a higher marketable fruit yield per crop. The total soluble solids were cultivar-specific, with the Brix level of Chevello changing with planting time while those of Chevi remained constant. The study identifies the potential for an alternative protected cropping structure, i.e., the modified multi-span polytunnels. The technical feasibility and affordability of the alternative protected cropping structure is also discussed.

1. Introduction

Ensuring food security for the burgeoning world population, which has been steadily increasing—exceeding 8 billion people in 2022—and is forecasted to reach 10 billion people by 2058, has become more challenging than ever before [1]. In semi-arid climates characterised by high temperatures, high solar radiation in spring and summer, low annual rainfall, and low relative humidity, extreme weather events affect growth, development, and yield, as well as quality of vegetable crops [2]. These extreme conditions are likely to be exacerbated in the future by climate change. It is predicted that, by 2050, global temperatures will increase by 1.5 °C, with more frequent and intense heat waves, storms, and droughts, along with more unpredictable rainfall, making agricultural production in semi-arid regions even more difficult [3].

Carnarvon is an important horticultural district which is located approximately 900 km north of Perth, the capital of Western Australia which has semi-arid weather conditions [2]. Capsicum is the second most important vegetable crop after tomato in the district, with approximately 3,700 tonnes produced annually and having a total farm gate value of AUD 13.5 million [4]. High temperatures and excessive sunlight, low air humidity, and high wind speed in spring and summer are the main limitations to the attainment of stable, high yield, and high quality capsicums in Carnarvon [5]. The shade net house is a popular protected cropping structure currently used in this area to cope with the adverse weather conditions. This structure type is made of wooden poles, steel cables, and netting materials covering the roof and all four sides, which can protect crops from high radiation and strong wind. Capsicums are usually planted from March (early autumn) to May (late autumn) and the harvest usually ends in October (spring) of the same year under this protected cultivation. However, the microclimate inside the shade net house is not always optimal for the early and late parts of the growing season, resulting in a shorter harvest window and lower overall production. Previous studies have shown that the shade net houses only reduced the excessive sunlight and wind speeds to a certain extent. Since the roofing is permeable, they cannot shelter the crops from unwanted rain, morning dew, and dust. Control of these factors is very important, as they can lead to the spread and proliferation of airborne fungal diseases that seriously affect vegetable crops [6,7,8].

Using protected cropping structures and altering the planting dates are among some of the strategies that the vegetable industry uses to cope with climate variability [9]. Among the protective structures suitable for Carnarvon weather conditions is the RRPS, which consists of a retractable roof, a retractable insect net below the roof, retractable curtains at all four sides, and a fogging system. The fine-tuned operation of these elements provides a modifiable growing environment that can protect crops from unpredictable weather extremes [5]. Our previous studies demonstrated that capsicum crops grown in the RRPS outperformed those in the open field (OF) for fruit yield and quality owing to the ability to modify the internal growing environment [5]. However, the optimum planting dates under the RRPS remain to be identified for crop yield per unit area improvement, as well as the extension of the supply season to target the high-priced market windows, increasing the return on investment. Moreover, the concerns of the local vegetable industry in Carnarvon on the high capital cost and the potential return on the investment of the RRPS are major factors with respect to grower adoption. Further research to determine an alternative protected cropping structure which is technically suitable and more affordable to local vegetable growers in Carnarvon is also needed.

This industry-focused research—part of the protected cropping research programme for vegetable crops in Western Australia, administered by the Department of Primary Industries and Regional Development (DPIRD)—was conducted in the Carnarvon and Geraldton regions (Figure 1). The aim of this study was to: (i) evaluate whether the current RRPS, with more advanced features than the shade net houses, can improve capsicum production in the early and late parts of the growing season, and (ii) explore the potential for an alternative protected cropping structure that is more affordable and able to grow vegetable crops productively under Carnarvon’s semi-arid weather conditions.

2. Materials and Methods

2.1. Experimental Design and Operation of the RRPS to Optimise Internal Climatic Conditions

The trial was conducted in the RRPS at the Carnarvon Research Station, DPIRD (24°51′18.00″ S latitude, 113°43′46.33″ E longitude) in 2022 (Figure 1A). According to the multi-year statistics of the Australian Bureau of Meteorology, the annual means of maximum and minimum temperatures were 27.4 °C and 17.2 °C, respectively, and the average annual rainfall was 221.6 mm. The soil inside the RRPS was red silty loam soil and was prepared before planting, as previously described by Nguyen et al. [5]. In brief, the raised beds of 60 cm wide and 10 cm high, spaced 1.2 m, were created by the tractor. The polyethylene plastic mulch and drip irrigation tape were placed together using a plastic layer attached to the rear of the tractor during the bed forming. The trial was a split-plot design with three replicates, where planting times are regarded as main plots and varieties as subplots. Each experimental plot was 60 cm wide × 10.5 m long. There were three planting dates, i.e., 28 February 2022, 8 April 2022, and 23 May 2022, which are designated hereafter as P1, P2, and P3, respectively. The trial was completed on 8 December 2022 for all planting dates. Two capsicum varieties, namely Chevi and Chevello, which were kindly provided by HM Clause Pacific seed company (HM Clause Pacific Pty. Ltd., Templestowe Lower, VIC, Australia), were used in this experiment. Chevi is an early maturing variety characterized by an excellent fruit set uniformity and premium fruit quality, with a bright red colouration and medium-sized blocky fruit. Chevello is also an early maturing variety exhibiting a large uniform fruit size, bright yellow colouration, and very firm and well-shaped fruit. Both varieties have intermediate resistance to the tomato spotted wilt virus, which is one of the critical viruses transmitted by western flower thrips in Carnarvon. Five-week-old capsicum seedlings were planted in double zigzag rows on the raised beds at a distance of 25 cm between rows and 30 cm within each row to obtain the final density of 35,088 plants per hectare.

The fertilising programmes used in the experiment for the three planting dates are presented in Supplementary Tables S1–S3. The following fertilisers were used in the trial: monoammonium phosphate, Poly-feed GG low K, potassium sulphate, potassium nitrate, magnesium nitrate, calcium nitrate, and Gripper trace 8. The gypsum basal fertilisation was applied before laying the plastic mulch, and the weekly fertigation began one week after the planting. The pruning and training of capsicum plants to the horizontal trellis were carried out fortnightly, commencing three weeks after planting and following the method described previously [5]. Insect pests and diseases were monitored every second day and controlled by using appropriate methods such as biologicals and chemicals.

The detailed strategic operation of the RRPS was carried out as previously described by Nguyen et al. [5]. Briefly, the roof and all side curtains were fully opened at 7:00 a.m., when the external air temperature was >14 °C, and were closed at 6:00 p.m., when the external air temperature was <18 °C. In the daytime, if the black plate temperature was above 32 °C, the roof would be opened 40% and further reduced to 10% if the internal relative humidity was less than 40%. Simultaneously, all side curtains would be fully closed, while the insect net would be fully opened. The fogging system was activated to increase the internal humidity when the internal temperature was >28 °C and the internal relative humidity was <60%.

2.2. Installation of Temperature, Humidity, PAR, and Wind Sensors

The weather conditions were monitored during the trial period at two sites, i.e., Carnarvon Research Station and a grower’s property in Geraldton (28°51′25.31″ S latitude, 114°43′37.84″ E longitude) (Figure 1B). The Geraldton region has a Mediterranean climate, but it is largely influenced by a semi-arid climate [10]. The annual means of maximum and minimum temperatures were 24.7 °C and 14.4 °C, respectively, and the average annual rainfall was 439 mm (Australian Bureau of Meteorology).

Two micro weather stations, Watchdog model 1450 from Spectrum Technologies, Inc. (distributed by John Morris Group, Bentley, WA, Australia) described previously by Nguyen et al. [2], were installed inside and outside the RRPS at the Carnarvon Research Station. In brief, the weather station comprises temperature, relative humidity (RH), and quantum light (PAR) sensors. All measurements were set to automatically log every 15 min. Two wind sensors with wind loggers (LeWL_Pro) from Logic Energy Ltd., Kilmarnock, UK (distributed by Instrument Choice Pty Ltd., Dry Creek, Adelaide, Australia) were installed inside the RRPS and in the OF. Another two identical sets of micro weather stations and wind sensors with loggers were installed inside the multi-span polytunnels and OF at the grower’s property in Geraldton. The weather data was downloaded monthly and used to calculate daily light integral (DLI), hourly mean temperature, and hourly mean RH following the method in Nguyen et al. [2].

2.3. Plant Height Measurement at Crop Establishment

The height of the capsicum plants was determined following the method described previously by Nguyen et al. [5]. Ten plants per experimental plot were selected for height determination. After planting, plant heights were measured weekly for four consecutive weeks for all three planting dates.

2.4. Marketable Fruit Number and Yield

Mature fruits from each experimental plot were harvested at 10–14-day intervals. The fruits with reasonably good size, free from marks, blemishes, and insect damage as described elsewhere [5] were graded as marketable, while the rest were categorised as rejects. Fruit number, marketable fruit weight, average fruit weight, and reject percentage were pooled from multiple harvests for each experimental plot. The marketable fruit yield per ha was calculated by using the total marketable fruit weight and the planted area following the method in Nguyen et al. [5].

2.5. Total Soluble Solids

The juice of three representative capsicum fruit per experimental plot was extracted by an extractor. The total soluble solids (°Brix) of the juice were determined by using a digital refractometer Atago RP-1 (Atago Co., Ltd., Kyoto, Japan), as described elsewhere [5].

2.6. Statistical Analyses

Analysis of variance (ANOVA) was performed to analyse the planting date effects, the varietal effects, and their interaction using GENSTAT statistical software version 21 (VSN International Ltd., Hemel Hempstead, UK). Means were compared at 5% probability according to the least significant difference (l.s.d) test.

3. Results

3.1. The Climatic Conditions inside the RRPS for the First Month after Each Planting Date

The planting dates were in late February, late April, and early May, but crops were completed on the same date in early December. To understand how the planting date affected the establishment, growth, and fruit yield of capsicum varieties, we collected the climatic data during the first month after planting for all three planting dates (Figure 2). The average light intensity in the first month of P1 was much higher than that of P2 and P3 (Figure 2A). Mean DLI of P1 during the first month was 18.61 mol m⁻² d⁻¹, while the mean DLIs for P2 and P3 during the first month were 14.85 mol m⁻² d⁻¹ and 14.19 mol m⁻² d⁻¹, respectively.

Mean air temperature during the first month of P1 was higher than that of P2 and was lowest for P3, with the average day—night temperature of P1, P2, and P3 being 27.5 °C, 23.5 °C, and 20 °C, respectively (Figure 2B). Mean daytime temperature (7:00 a.m.–6:00 p.m.) for P1, P2, and P3 was 29.8 °C, 26.1 °C, and 22.3 °C, respectively. The P1 treatment experienced fairly high average daytime temperatures, with the temperature rising above 36 °C on several days; conversely, P3 experienced lower average temperatures, the latter often reaching only 22–25 °C. Meanwhile, the average nighttime temperature (7:00 p.m.–6:00 a.m. of the next day) for P1, P2, and P3 was 25.1 °C, 20.9 °C, and 17.7 °C, respectively. The lowest nighttime temperature for P1 was 20 °C, while P3 experienced the lowest temperature, i.e., 11 °C.

The mean nighttime RH (7:00 p.m.–6:00 a.m. of the next day) for P2 and P3 was higher than that for P1 (Figure 2C). Meanwhile, the mean nighttime RH (7:00 p.m.–6:00 a.m. of the following day) for P1, P2, and P3 was 72.2%, 73.0%, and 81.2%, respectively (Figure 2C). The average daytime RH (7:00 a.m.–6:00 p.m.) for P1, P2, and P3 was 58.5%, 55.9%, and 65.5%, respectively.

3.2. Crop Growth

To understand how the weather conditions during the first month after planting affected crop growth, weekly plant height observations were collected for the first four weeks. The data showed that the height increase of P1 was the fastest, followed by P2; P3 experienced the slowest increase in height (p < 0.05, Figure 3). The average height across the three planting times for both varieties after four weeks was 45.6 cm, 36.1 cm, and 29.8 cm, respectively (p < 0.01, Figure 3). The average weekly growth rate of the three plantings was 8.6 cm, 6.6 cm, and 4.6 cm, respectively, for both varieties. Of the two varieties, Chevi experienced a slightly faster growth rate than Chevello, with an average weekly growth rate of 6.9 cm and 6.3 cm, respectively.

3.3. Marketable Fruit Yield and Yield Attributes

In the 2022 cropping season, the first harvest dates of the three plantings were 17 May (P1–2.5 months after planting), 7 July (P2–3 months after planting), and 15 September (P3–3.5 months after planting). The total number of pickings for the whole growing season of the three plantings were 22, 16, and 9, respectively.

The data from the full growing cycle showed that there was no interaction between the planting date and variety (

Table 1. Marketable fruit yields, fruit per plant, fruit weight, and reject rate of capsicum varieties at different planting times. Means of the main effects in the same row (or column) that include a common letter are not significantly different at 5%. P, planting; V, variety. The p values indicate the statistical significance at 5%; s.e.d, standard error of differences of the means; l.s.d, least significant difference.

Variety	Marketable Fruit Yield (t ha−1)				Fruit Per Plant				Average Fruit Weight (g)				Reject Rate (%)
	P1	P2	P3	Mean	P1	P2	P3	Mean	P1	P2	P3	Mean	P1	P2	P3	Mean
Chevello	103.4	70.4	33.3	69.0 a	16.0	10.5	4.8	10.4 a	184.5	190.2	198.9	191.2 a	20.7	24.6	36.7	27.3 a
Chevi	101.8	74.7	38.9	71.8 b	14.4	9.6	4.9	9.6 a	201.9	220.9	225.0	215.9 b	20.8	23.7	27.7	24.0 b
Mean	102.6 a	72.5 b	36.1 c		15.2 a	10.1 b	4.8 c		193.2 a	205.5 b	212.0 b		20.7 a	24.2 a	32.2 b
ANOVA	P	V	P × V		P	V	P × V		P	V	P × V		P	V	P × V
s.e.d	4.50	3.54	6.25		0.43	0.45	0.7		5.05	1.87	5.54		2.54	1.18	2.92
p	<0.001	0.47	0.69		<0.001	0.136	0.34		0.026	<0.001	0.067		0.01	0.032	0.038
l.s.d (p = 0.05)	12.50	-	-		1.05	-	-		12.35	4.58	-		6.22	2.90	6.56

; p = 0.69). On average, P3 resulted in the lowest marketable fruit yield (36.1 t ha⁻¹), which was about half of the marketable fruit yield attained by P2 (72.5 t ha⁻¹) and one third of the marketable fruit yield achieved by P1 (102.6 t ha⁻¹) (Table 1; p < 0.001). The marketable fruit yields were similar between the two capsicum varieties for all planting dates (Table 1; p = 0.47). The difference in fruit yield among the planting dates was mainly determined by the average number of fruits per plant (Table 1; p < 0.001) and the rejection rate (Table 1; p = 0.01). Interestingly, P1 attained the highest average fruit yield, but the mean fruit weight was significantly smaller than that of P2 and P3 (Table 1; p = 0.026). There was no statistically significant difference between the two capsicum varieties in the mean number of fruits per plant (Table 1; p = 0.136). However, there was a significant difference between the two capsicum varieties with respect to the mean fruit weight (Table 1; p < 0.001) and the reject rate (Table 1; p = 0.032). Chevi was characterized by a larger average fruit weight (215.9 g) than Chevello (191.2 g), as well as resulting in a lower average reject rate (24%) compared to Chevello (27.3%).

To find out the influence of seasonal factors on fruit yield and yield contributing components, we compared the harvest data from the first 6 months after planting for all three planting dates (

Table 2. Marketable fruit yields, fruit per plant, fruit weight, and reject rate of capsicum varieties 6 months after planting. Means of the main effects in the same row (or column) that include a common letter are not significantly different at 5%. P, planting; V, variety. The p values indicate the statistical significance at 5%; s.e.d, standard error of differences of the means; l.s.d, least significant difference.

Variety	Marketable Fruit Yield (t.ha−1)				Fruit Per Plant				Fruit Weight (g)				Reject Rate
	P1	P2	P3	Mean	P1	P2	P3	Mean	P1	P2	P3	Mean	P1	P2	P3	Mean
Chevello	50.5	41.7	30.9	41.0 a	7.5	6.1	4.3	6.0 a	191.7	195.3	202.7	196.6 a	7.7	8.7	35.0	17.1 a
Chevi	53.8	45.2	36.4	45.1 b	7.2	5.4	4.5	5.7 a	213.1	240.6	232.3	228.7 b	9.2	8.9	26.2	14.8 b
Mean	52.1 a	43.5 ab	33.6 b		7.3 a	5.7 b	4.4 c		202.4 a	218.0 b	217.5 b		8.4 a	8.8 a	30.6 b
ANOVA	P	V	P × V		P	V	P × V		P	V	P × V		P	V	P × V
s.e.d	4.02	1.03	4.21		0.45	0.17	0.49		5.22	2.63	6.13		1.74	0.93	2.08
p	0.025	0.007	0.644		0.007	0.113	0.182		0.066	<0.001	0.026		<0.01	0.043	0.007
l.s.d (p = 0.05)	11.15	2.51	-		1.25	-	-		-	6.43	14.52		4.84	2.26	4.89

). The results showed that, similarly to the full growing cycle, there was no interaction between planting date and variety (Table 2; p = 0.644). Still, P1 attained the highest marketable fruit yield (52.1 t ha⁻¹), followed by P2 (43.5 t ha⁻¹), with P3 attaining the lowest marketable fruit yield (33.6 t ha⁻¹). The difference in the marketable fruit yields between the plantings was largely determined by the average number of fruits per plant (Table 1; p = 0.007). Intriguingly, the reject rate during the first 6 months of P3 (30.6%) was much higher than that for P1 (8.4%) or P2 (8.8%) alone.

3.4. Total Soluble Solids

There was no interaction between planting time and variety for the total soluble solids (

Table 3. Total soluble solids of capsicum varieties. Means of the main effects in the same row (or column) that include a common letter are not significantly different at 5%. P, planting time; V, variety. The p values indicate the statistical significance at 5%; s.e.d, standard error of differences of the means; l.s.d, least significant difference.

Variety	Total Soluble Solids (°Brix)
	P1	P2	P3	Mean
Chevello	5.9	5.5	6.5	5.9 a
Chevi	6.6	6.6	6.6	6.6 b
Mean	6.3 a	6.0 b	6.6 c
ANOVA	P	V	P × V
s.e.d	0.11	0.21	0.28
p	0.009	0.018	0.223
l.s.d (p = 0.05)	0.28	0.51	-

; p = 0.223). The planting time had no effect on Chevi’s °Brix, but it did affect Chevello’s °Brix (Table 3; p = 0.018). The effect of the planting time on °Brix was different for all three plantings. P3 had the highest °Brix (6.5), followed by P1 (5.9) and P2, which had the lowest °Brix (5.5) (Table 3).

3.5. Internal and External Environmental Conditions of the Multi-Span Polytunnels in Geraldton

To investigate the variability in the climatic conditions between the Carnarvon and Geraldton regions, as well as the difference caused by protective structure types, we compared the weather conditions in the OF and in the RRPS in Carnarvon with those in the OF and in the multi-span polytunnels in Geraldton (Figure 1). We used the representative weather data from two months in 2022, i.e., March (early autumn) and July (winter), for the analysis (Figure 4).

In the OF, the light intensity in Carnarvon was overall higher than that in Geraldton for both early autumn and winter, with the mean DLIs in March being 51.8 mol m⁻² d⁻¹ and 48 mol m⁻² d⁻¹, respectively (Figure 4A), and 31.9 mol m⁻² d⁻¹ and 25.4 mol m⁻² d⁻¹, respectively, in July (Figure 4B). On average, the protected structures at both locations caused a reduction in light intensity of 40% in the early autumn and of 50% in the winter (Figure 4A,B). During early autumn, light intensity in the RRPS (18.4 mol m⁻² d⁻¹) was slightly lower than that in the multi-span polytunnels (20.5 mol m⁻² d⁻¹). However, the opposite was true in the winter, with the light in RRPS (15.5 mol m⁻² d⁻¹) being slightly higher than that in the multi-span polytunnels (14.2 mol m⁻² d⁻¹) (Figure 4A,B).

In the OF, Carnarvon experienced higher mean daytime and nighttime air temperatures than the site in Geraldton during the early autumn and winter (Figure 4C,D). In Geraldton, the average air temperature in the multi-span polytunnels was much higher than that in the OF during the day (approximately 3–5 °C), but they were only slightly higher at night (approximately 0.5–1 °C) (Figure 4C). During the early autumn, the mean air temperature in the RRPS was lower than that in the OF in Carnarvon during the daytime (about 3–5 °C), but they were similar at night (Figure 4D). During winter, the mean daytime and nighttime air temperatures of the RRPS were slightly higher than those experienced in the OF in Carnarvon (about 0.5–1 °C) (Figure 4D).

Overall, the relative humidity in the OF in Carnarvon and Geraldton was low during the day in early autumn (approximately 25–35%) and elevated at night (approximately 70–80%) (Figure 4E). There was not much difference in the mean daytime and nighttime relative humidity between Carnarvon and Geraldton in OF during the early autumn (Figure 4E). In the winter, the average relative humidity in the OF in Carnarvon was approximately 15% lower than that experienced in Geraldton (Figure 4F). Interestingly, the average relative humidity in the multi-span polytunnel was approximately 5% higher than that experienced in the OF in Geraldton during early autumn, but it was approximately 9% lower in the winter. Meanwhile, the average air relative humidity in RRPS was always approximately 5% higher than that in the OF in Carnarvon for both winter and early autumn (Figure 4E,F).

The wind speed in the OF in Geraldton was higher than that in Carnarvon, with the mean daily wind speed reaching 10.78 km h⁻¹ and 8.10 km h⁻¹, respectively, in early autumn (Figure 4G) and 6.10 km h⁻¹ and 3.63 km h⁻¹, respectively, in winter (Figure 4H). The protected structures, i.e., multi-span polytunnels and RRPS, significantly reduced the wind speed at both the Geraldton and Carnarvon sites, with the mean daily wind speed reaching 0.93 km h⁻¹ and 0.74 km h⁻¹, respectively, in early autumn (Figure 4G) and 0.03 km h⁻¹ and 0.38 km h⁻¹,respectively, in winter (Figure 4H). Interestingly, the mean daily wind speed within the multi-span polytunnels was higher than that experienced in the RRPS in early autumn, although it was lower in the winter (Figure 4G,H). This seasonal discrepancy was probably due to the difference in the operation of the ventilation systems in the two protective structures.

4. Discussion

4.1. Effects of the Planting Dates on the Growth of Capsicum Varieties

This study aimed to provide an understanding of the seasonal effects on the growth and the marketable fruit yield of capsicum varieties grown in the protected cropping system RRPS in Carnarvon. In this manuscript, we have shown how the seasonal factors critically affected the crop growth, the timing, and the number of harvests, all of which determine the final marketable fruit yield. The research also showed that the final marketable fruit yield was highly dependent on the crop establishment and the early growth stage. Temperature and sunlight were the elements determining the growth rate of capsicums, the number of fruits per plant, the reject rate, and the total soluble solids.

The yield in the first 6 months of P1 was higher than that of P2 and P3 due to early planting, with the temperature and the light intensity in the first month of P1 being much higher than that of P2 and P3 (Figure 2A, Table 2 and Table 3). The relatively high temperatures of P1 and P2 at the establishment stage resulted in better plant growth (Figure 3) and a shorter time interval between planting and the first harvest. In contrast, P3 experienced the longest time interval between planting and the first harvest due to low temperatures and low light intensity in the first month, so the plants grew slowly and had fewer fruits per plant. Previous studies showed that flowering in tomato and capsicum plants was dependent on growing degree days, which were calculated as the sum of the cumulative temperatures required to induce flowering [11,12]. The effects of the planting time on total soluble solids were cultivar-specific. The Brix of the Chevello variety varied with planting time, while that of the Chevi variety remained constant (Table 3). These results indicate that breeding programmes should target the quality traits, in addition to the yield, by selecting suitable varieties for protected cultivation to obtain a constant produce quality.

The optimum temperature for the growth and development of capsicum plants is around 20–25 °C [13]. Since seedlings were planted during a hot month (late summer), the harvests from the first 6 months of P1 had a greater number of pickings and a higher marketable fruit yield than the other two planting dates. P3 was planted during a period when the temperatures were lower (winter), causing the plants to grow slowly, the time period between planting and first picking to be longer, the number of pickings to be lower, and the marketable fruit yield to be lower.

Our previous research showed that weather conditions have an extremely important influence on the growth, development, and yields of fruiting vegetables such as tomato, eggplant, and capsicum [2,5]. Unfavourable weather conditions such as extreme heat, low light levels, biological factors, infestations of insect pests, and diseases will lead to a drop in flower and fruit production, resulting in low fruit yields [14,15]. This might be the reason why P3 had a much lower number of fruits per plant than P1 and P2. Suboptimal weather conditions in the early growth stages also led to weaker plant growth and high rejection rates (Table 2). This study is also supported by previous research which found that late planting dates are at higher risk of succumbing to other abiotic and biotic threats [16].

4.2. The RRPS Supports a Longer Growing Season to Increase the Overall Marketable Fruit Yield and Extends the Supply Window

In Carnarvon, capsicums are usually planted between March and May, and harvesting is usually finished by October of the same year in shade net houses. However, the internal microenvironment in the shade net houses is sub-optimal for the early planting in February and at the end of the season in October due to excessive heat and low humidity during the day, resulting in a shorter harvest window and overall lower production. The results from our study showed that the longer the growing season, the more abundant the harvests, leading to a higher marketable fruit yield. The average marketable fruit yield of the capsicums under the full growing season from February to December was more than 100 tonnes ha⁻¹ (Table 3) compared to an average of 70 tonnes ha⁻¹ of capsicums grown under growers’ shade net houses. Most vegetable growers using shade net houses in Carnarvon plant their crop in March and complete all harvests by the end of October, as the shade net house is only able to reduce the solar radiation and the wind speed to a certain extent [2]. This cropping pattern significantly reduces crop yields and competitiveness in the market, as well as overall profitability. However, this limitation could be mostly mitigated by using a modern RRPS system equipped with a suitable shade screen and a fogging system [5]. The under-roof shade screens help reduce excess light intensity (Figure 4A), while the fogging system can cool down the internal air temperature (Figure 3C) while simultaneously increasing air humidity (Figure 4E). Thus, the RRPS allows the planting of capsicum to be conducted as early as February, enabling early harvests at the beginning of May for sales in the Perth market, bringing greater profit to growers. Past studies revealed that the capsicum price in the Perth market is usually elevated in November and December due to the extreme sunlight and the strong wind which can result in sunburn and wind damage to fruit, reducing the supply volume from Carnarvon [17]. Extending the season further into January would target high priced markets which would improve grower profitability. Climate change predictions for Carnarvon include hotter weather and increased extreme weather events such as cyclones and more summer rainfall. The RRPS is a flexible protected cropping structure that is well suited to hot climates, as heat can be rapidly dispersed by opening the roof which, in turn, can be closed when rainfall is imminent. The RRPS is suited to climates which experience cyclones, as the roof and sides can be retracted preventing damage to the structure. Overall, the RRPS can potentially alter and improve the existing vegetable cropping system in Carnarvon under the current weather conditions and climate change.

4.3. Potential for a Modified Multi-Span Polytunnel Structure for Carnarvon

Although it possesses many advanced features compared to the shade net houses currently used in Carnarvon, the RRPS is still considered less attractive by many growers because of its high initial investment cost [5]. Previous studies demonstrated that identifying a low-cost but suitable protected cropping option is critical for growers. If the payback period was short, the capital investment risk would be reduced, and thus the newly introduced technology would be more easily adopted by the growers [18]. Therefore, it is essential to identify a protected cropping system that is as technically sound as the RRPS, but more affordable.

Located approximately 475 kilometres south of the Carnarvon region, Geraldton is also an important horticultural district in the mid-west region of Western Australia. Geraldton farmers grow fruiting vegetables such as tomato, capsicum, and cucumber in unheated and naturally ventilated multi-span polytunnels (Figure 5A,B), producing crops with a total value of about $19.4 million per annum [19]. Along with Carnarvon, Geraldton provides fresh winter produce when the southern regions cannot due to cold weather. The multi-span polytunnel structure is low-tech, but can effectively protect crops from unfavourable weather conditions such as unwanted rain, morning dew, dust, and strong winds. Moreover, this structure can also efficiently prevent the entry of vectoring insects that transmit serious viral diseases, such as thrips, whiteflies, and aphids, by using suitable covering materials [20,21,22]. Virus infections o can result in severe viral diseases such as the tomato spotted wilt and the cucumber mosaic virus, both of which devastate vegetable production in Carnarvon [23].

Comparison of the climatic conditions between Geraldton and Carnarvon revealed that Geraldton shared some similarities with respect to the light conditions. However, temperatures, humidities, and wind speeds were different (Figure 4). The multi-span polytunnels’ temperatures were still high in early autumn, leaving growers with fewer options to grow profitable summer crops (Figure 4C). The high temperatures in multi-span polytunnels were due to the greenhouse effect and poor ventilation, since the roofing is completely covered by plastics [24]. Using the fogging system to bring down the temperatures in the summer without a proper ventilation would create an excessively high internal RH, causing the development of fungal diseases, e.g., powdery mildew (Ly per. comm.). The excessive heat will also reduce the crop’s calcium uptake, leading to calcium deficiency in fruit and occurrence of blossom end rot [25,26].

To ensure that the new multi-span polytunnels can be effectively used under Carnarvon’s climatic conditions, a prerequisite is that the internal environmental factors, i.e., temperature, humidity, and light, must be well controlled [27]. The multi-span polytunnels used in Geraldton can potentially be used in Carnarvon with some modifications. The structural details of the multi-span polytunnel type for warm regions such as Carnarvon have been extensively discussed elsewhere [28]. This structure type should have a higher clearance and curved roof where the arches overlap to create a roof vent. These vertical openings in the roof vent would allow for a more effective escape of the hot air through passive ventilation as we observed in the RRPS. The roof vent could be fitted with an insect exclusion net. This design prevents the entry of rain and morning dew, whilst providing continuous ventilation. The modified multi-span polytunnels could be installed with an under-roof retractable shade screen and a fogging system similar to that of the RRPS, while all four sides could be fitted with fixed insect exclusion nets and roll-up plastic curtains. Although the temperature is high during the summer, the wind is usually strong in Carnarvon during this time of the year (Figure 4G). The strong wind will aid the modified multi-span polytunnels in discharging the internal hot air and excessive humidity from fogging even more effectively. Khapte et al. [29] proved that tomato crops grown in the naturally ventilated polytunnels performed better than insect-proof net houses and shade net houses in semi-arid regions, owing to superior plant growth, marketable fruit yields, and fruit quality.

This modified multi-span polytunnel type would be more affordable than the RRPS (Pham per. comm.), while most of the advanced features of the RRPS would be retained. This system would allow Carnarvon farmers to grow a variety of vegetables over a considerable portion of the year because the under-roof retractable shade screen can be used to modify the level of radiation entering the structure and the high roof with ventilation will assist with cooling.

More variability in weather conditions due to climate change will pose substantial challenges to the vegetable industry in the future, causing unpredictable weather conditions, higher temperatures, frequent heatwaves, and high insect pest and disease pressure. Protected cropping will become an indispensable production system for the local vegetable growers in semi-arid regions to mitigate the negative impacts and risks associated with climate change. Therefore, modified multi-span polytunnels with more effective cooling and ventilation capacities possibly represent a valid future choice for vegetable growers in Carnarvon who use protected cultivation to address the challenges of climate change.

5. Conclusions

In this study, we provided experimental evidence to demonstrate the effects of the planting date on crop establishment, growth, and timing and number of harvests, all of which determine the final marketable fruit yield. The results demonstrate that a capsicum crop can be grown in late summer—with harvesting completed in late spring—in Carnarvon by using the RRPS, without compromising the marketable fruit yield and quality. The RRPS effectively mitigated the high temperature, excessive sunlight, and the low RH to provide an optimised internal microenvironment for better crop establishment and extended harvest in late spring, leading to a higher marketable fruit yield per crop. The study also identified the potential for an alternative protected cropping structure, i.e., the modified multi-span polytunnels that retains important features of the RRPS but is more affordable.

It is recommended that further investigations are conducted into the modified structure to confirm its reliability under the weather conditions specific to Carnarvon. Future trials should utilise elite capsicum cultivars and other vegetable varieties that are tolerant to biotic and abiotic conditions together with the optimum fertilisation and irrigation to improve nutrient use efficiency and water use efficiency under this protected cropping system to sustain productivity, storage life, and premium quality produce.