9

Response of fiber hemp (Cannabis sativa L.) to varying irrigation regimes

Shaun Lisson¹ and Neville Mendham²

¹CSIRO, Davies Laboratory, University Drive, Aitkenvale, Queensland 4814, Australia

²Department of Agricultural Science, University of Tasmania, Hobart, Tasmania 7001, Australia

        Lisson, Shaun and Neville Mendham 1998. Response of fiber hemp (Cannabis sativa L.) to varying irrigation regimes. Journal of the International Hemp Association 5(1): 9-15. This paper reports on the findings from a hemp (Cannabis sativa L.) irrigation trial conducted on a krasnozem soil in northwest Tasmania, Australia in 1995/96. Studies were made of crop growth, development and water use responses to rainfed and irrigated treatments with allowable soil water deficits of 30, 60, 90 and 120 mm. At final harvest, the stem yield and total dry matter of the irrigated treatments were not significantly different, but each yielded more than the rainfed treatment. The percentage of bark in the stem appeared to decline with deficit to irrigation. When translated into bark yield at final harvest, the 30 mm and 60 mm deficit treatments yielded significantly more bark than the remaining treatments. Similarly, the 90 mm and 120 mm treatments were higher yielding than the rainfed treatment. Soil water extraction was primarily confined to the upper 80-90 cm of the soil profile, with little extraction at greater depths. In the rainfed treatment, there was evidence of minor extraction down to at least 140 cm. Water consumption (to 90 cm) ranged from 359 mm for the rainfed treatment, to 535 mm for the 60 mm treatment. Water use efficiency ranged from 2.4 g/kg for the 90 mm treatment to 3.1 g/kg for the 120 mm treatment.

Introduction
        Literature recommendations for the water requirements of fiber hemp are sparse and somewhat ambiguous. In the Ukraine, optimum yields are achieved with 250-280 mm of rainfall during the growing season (P. Goloborodko. pers. comm. 1994). Begg and Buller (1995) suggest a total available water requirement of approximately 600 mm. Van Dam (1995) reports that hemp requires rainfall of at least 650 mm per year.
        Preliminary field trials with hemp in Tasmania indicated that irrigation would be essential in order to overcome deficiencies in both the distribution and the total amount of rainfall over the summer months (Lisson 1998). Areas of a hemp trial conducted in 1994/95 that received insufficient water were severely stunted and gave poor stem and fiber yields in comparison to well-watered sites. Furthermore, a growth model developed by Hackett (1991) identified rainfall shortages as a major limitation to hemp growth in three separate locations across southern Australia.
        The purpose of this trial was to examine the effect of a range of allowable soil water deficits on the yield and growth responses of hemp (cv. Kompolti). Sequential harvest data enabled examination of the ability of fiber hemp to recover from periods of water stress. Regular measurements of soil water content with a neutron moisture meter were used to develop soil water extraction profiles and to calculate seasonal water consumption and water use efficiency.

Materials and methods
Design & treatments-
        A randomized complete block design was employed with four replicates. Each plot was 1.6 m wide and 10 m long. A buffer zone (cv. Kompolti) of one plot width was sown around the perimeter of the trial area.
        The five irrigation treatments covered a range of soil moisture deficits including 30 mm (I₃₀), 60 mm (I₆₀), 90 mm (I₉₀), 120 mm (I₁₂₀) and a rainfed treatment (I₀). The number and dates of irrigations are shown in Table 1. Deficits were based on cumulative, rainfall-adjusted Class A pan evaporation measurements made within 500 m of the trial site. Upon reaching the treatment deficit, each of the four replicate plots was returned to field capacity.
        Total rainfall for the duration of the trial was approximately 60 mm above the average (38 year) rainfall total for the same period (Figure 1). The distribution of rainfall was less favorable, with lower than average rainfall totals for the months of November and especially December. Over twice the average rainfall was received in January.

Figure 1. (a) Total rainfall and Class A pan evaporation for five day periods (days after sowing and date (bracketed) along the X axis) from 20/9/95 to 7/2/96. (b) Monthly average rainfall totals from 1958-1996 and 1996-97 season monthly totals.

Cultural methods-
        Seed of the cultivar 'Kompolti' was sown to a depth of approximately 3-4 cm, using a 10 row cone seeder with 13.5 cm row spacing and 1.6 m plot centers. The entire trial area was sown at a seeding rate of 80 kg/ha, to establish a plant density of 150 plants/m 2 .
        The trial was sown on September 26, 1995 at the Forthside Research Station (41°10’S, 146°40’E) on a krasnozem soil. This soil is classified as a reddish brown, strongly structured, gradational, clay loam to clay soil. The amount of useful water held in these soils types at field capacity down to a depth of 70 cm, is typically moderate at 10-15% (Bridge and Bell 1994, Chilvers 1996). Leeper and Uren (1993) report that these soils are susceptible to drought but their structure is such that roots are able to penetrate deep into the profile to reach greater reserves of water.
        Irrigation was applied with ‘Netafim’ flexible polyethylene drip tubing operated at a water pressure of 100 kPa and a delivery rate of 1.75 liters per hour per drip point. Four lengths of the flexible tubing were positioned along each plot at 30 cm spacing. Each set of treatment replicates were fed by a 19 mm polyethylene pipe which, in turn, was connected to a common 25 mm polyethylene pipe via an isolating tap.
        Fertilizer rates were based on a soil test and literature recommendations for hemp production (100-130 kg N/ha, 35-50 kg P/ha and 110-140 kg K/ha). A blend of 9:14:17 NPK fertilizer (350 kg/ha) and zinc sulfite monohydrate (30 kg/ha) was incorporated into the top 15-20 cm of the soil profile just prior to drilling. A further top dressing of nitrogen (40 kg/ha) fertilizer was applied 4-6 weeks after sowing.
        The trial area was hand weeded at ten weeks. Thereafter, weed growth was adequately suppressed by the hemp crop.
        Stem rot (Sclerotinia sclerotiorum) was apparent on isolated plants about the time of flowering but losses were not regarded as significant.

Data collection and analysis-
        Sequential harvests of 0.5 m 2 were collected at bi-weekly intervals, allowing 0.5 m of plot length as a buffer between each sample. The final harvest of 2.0 m 2 was collected at late flowering (Table 2). Dead and senesced leaf matter along with severely suppressed living plants (a result of ‘self-thinning’) were not included in harvested samples.
        Total above-ground fresh weight and plant counts were measured for each sample. A random selection of 20 plants was then set aside for partition analysis and measurement of stem length. Average leaf area (mm 2 ) was measured using a Paton/CSIRO Electronic Planimeter and was based on a cumulative sample of approximately 40 leaves. Stem length was measured from ground level to the top of the fruiting zone. Diameter was measured at the base of the stem. Node counts included the cotyledonary node. Paired nodes, with alternate leaves appearing after flowering were regarded as single nodes. Partitioned sub-samples were then oven-dried at 70°C for 48 hours to determine total dry matter and stem yield.
        At final harvest, the proportion of bark in the stem was determined from stem sections. Starting from about 15 cm above ground level, sections of approximately 30 cm in length were taken from 15 female plants. The bark was peeled away by hand and the relative proportions calculated from oven dry weights of each separated fraction.
        A neutron moisture meter (NMM) was used to

measure changes in percent volumetric soil moisture content with depth. One aluminum recording tube was buried to a depth of 1.5 m in three replicate plots of each treatment. Moisture readings were made every 3-4 days and averaged across the three replicates. The data were subsequently processed using a software package entitled ‘The Probe’, developed by Neutron Probe Services Pty. Ltd.
        Analysis of variance tests were performed using Systat 5.2.1 software. Means were compared using the Fisher LSD test with significance for P values less than 0.05.

Results
        Final harvest results are listed in Table 3.

Figure 2. Time responce plot for total dry matter. The bars correspond to Fisher LSD values (P<0.05) for the final three harvests.

Figure 3. Time responce plot for stem yield. The bars correspond to Fisher LSD values (P<0.05) for the final three harvests.

Yield and yield components-
        Significant treatment differences were apparent for total dry matter and stem yield in the fourth, fifth and final harvests (Table 3, Figures 2 and 3).
        Significant treatment differences were first apparent at Harvest 4 (day 98), with the total and stem dry matter yield of I₃₀ and I₆₀ exceeding those of the remaining treatments. There was an apparent delay in the response of I₉₀ (1st irrigation on day 69) and I₁₂₀ (1st irrigation on day 76) treatments to irrigation, with total dry matter and stem yield similar to that of the rainfed treatment at Harvest 4 (98 DAS, Figures 2 and 3).
        Following Harvest 4, the growth rates associated with I₉₀ and I₁₂₀ increased sharply, so that at Harvests 5 and 6, the stem yield and total dry matter of the irrigated treatments were not significantly different, but each yielded more than the rainfed treatment. This increase in growth rate coincided with substantial rainfalls and higher temperatures in early January.
        The percentage of bark in the stem appeared to decline with deficit to irrigation (Table 3). The I₃₀ and I₆₀ treatments had significantly higher bark percentages than the I₁₂₀ and rainfed treatments. Similarly, I₃₀ had a higher proportion than I₉₀.
        When translated into bark yield at final harvest, I₃₀ and I₆₀ yielded significantly more bark than the remaining treatments. Similarly, I₉₀ and I₁₂₀ were higher yielding than the rainfed treatment.

Figure 4. Time responce plot for leaf area index.

Leaf area production-
        There were no significant differences between treatments for leaf area index (LAI) or leaf yield at any of the harvests (Table 3, Figure 4). However, LAI for the 30 mm treatment was consistently greater than the rainfed treatment at all harvests. Furthermore, prior to flowering and the subsequent increase in leaf senescence (especially from male plants), LAI tended to decline with increased deficit.
        While there was no apparent trend in node number across the irrigated treatments at each harvest, the node number for the rainfed treatment was consistently less than that for the irrigated treatments (final harvest data shown in Table 3).

Figure 5. Time responce plot for stem length. The bars correspond to Fisher LSD values (P<0.05) for the final four harvests.

Stem length-
        Significant treatment differences were apparent for stem length in Harvests 3, 4, 5 and 6 (Table 3, Figure 5). The response to irrigation was readily apparent at Harvest 3, with I₃₀ and I₆₀ significantly taller than the recently irrigated I₉₀ and unirrigated I₁₂₀ and rainfed treatments. At Harvests 5 and 6, the irrigated treatments were not significantly different in height but were all taller than the rainfed treatment. This response resembles the response of total dry matter and stem yield.

Flowering-
        The time of flowering was taken as the 15th of January (112 days after sowing). There was no apparent difference between treatments.

Figure 6. Patterns of volumetric soil water content (%) with depth, at selected dates and at field capacity, for treatments: I₃₀ (a), I₆₀ (b), I₉₀ (c), I₁₂₀ (d), I₀ (e).

Soil water extraction-
        Figure 6 shows patterns of volumetric soil water content (VSW%) recorded just prior to irrigation or rain-fall events that returned the profile back to field capacity. These represent the maximum extent of crop water extraction for each irrigation treatment. The pattern corresponding to field capacity is also shown.
        Across all treatments, soil water extraction was primarily confined to the upper 80-90 cm of the soil profile, with little extraction at greater depths. In the rainfed treatment, there was evidence of minor extraction down to at least 140 cm. The distribution of water extraction in three zones down to 100 cm (0-35, 35-55 and 55-100), averaged over the patterns selected for Figure 6, are shown in Table 4. The decline in total water content for each depth interval is expressed as a percentage of the water content at field capacity. With the exception of the 60 mm treatment, extraction decreased with increasing soil depth. This decrease appeared greatest for the 30 mm treatment, with just under two thirds of the water extracted from the upper 35 cm of the profile. This may be attributable to a proliferation of roots in response to more frequent irrigation. In the other treatments, extraction appeared to shift into the 35 to 55 cm zone and, to a lesser extent, into the 55 to 100 cm zone.

Seasonal water received, water consumption and water use efficiency-
        The total amount of water received by each treatment is shown in Table 5. This total is the sum of rainfall and irrigation, discounted for losses through runoff when the amount of rainfall exceeded the prevailing deficit. This loss was most prevalent with the more frequently irrigated treatments, as reflected by the similar totals for the 30, 60 and 90 mm deficit treatments. The greatest amount of water was received by the 60 mm treatment and the lowest by the rainfed treatment.
        The change in water content between two NMM readings to a certain depth plus any intervening rainfall or irrigation amounts (discounted for runoff), gives a measure of the water consumed over that period. A sum of these periodic totals over the duration of the growing season gives the total water consumed. Seasonal water consumption to a depth of 90 cm is shown in Table 5. In the period between sowing and the first NMM reading, evapotranspiration was assumed equivalent to Class A pan evaporation. Dividing the mean stem dry weight by the total amount of water consumed through evapo-transpiration, gives a measure of the water use efficiency for stem production (Table 5).
        Similar water use efficiencies of about 3 g/kg were obtained for the 30 mm, 120 mm and rainfed treatments, with lower values for the 60 mm and 90 mm treatments. The high values for the two driest treatments may indicate that the roots reached a greater depth and were able to extract stored water lower in the profile. It may also indicate that water was used more economically due to stomatal closure during periods of water stress (Salisbury and Ross 1985).

Discussion
         Significant stem yield differences were not apparent for irrigation regimes based on refill to field capacity at deficits down to 120 mm. Seasonal water consumption for this regime was approximately 420 mm. Maximum bark yield was obtained from regimes based on a 60 mm deficit or less (water consumption of 535 mm). The significant decline in stem bark content with increasing deficit has previously been observed for kenaf (Hibiscus cannabinus L.) by Muchow and Wood (1980).
        Whilst the yield of the rainfed treatment was reasonable, lower stem yields might be expected under average weather conditions at Forthside, when the seasonal rainfall total is less and January rainfall is less than half that received in this trial. In the absence of the large rainfall event in January, the stem yield (Figure 3) for the rainfed treatment was likely to have only reached around 800-900 g/m 2 .
        Begg and Turner (1976) report that a reduction in leaf area is a major morphological consequence of water stress. This is primarily brought about through a reduction in cell division and more importantly, a lessening in the rate of cell enlargement. Although differences between leaf area indices across the five treatments were not significant, there was evidence of a declining trend with increasing deficit. Across the irrigated treatments, this was most likely attributable to a decline in area per leaf rather than a decline in leaf (node) or plant number. A decline in leaf number may have contributed to the decline in leaf area between the irrigated and rainfed treatments.
        Interestingly, there was a notable delay in the response of leaf area and dry matter production to the first irrigations for the 90 mm and 120 mm treatments. Eventually however, the growth rate increased sharply so that at Harvests 5 and 6, differences between the irrigated treatments for yield and total dry matter were insignificant. This increase coincided with an increase in leaf area index. Begg and Turner (1976) report that a delay in the opening of stomata and the recovery of photosynthesis following alleviation of water stress, has been observed in a number of other species. Increases in growth rate following watering of previously stressed plants have been reported for tomato (Gates 1955) and kenaf (Muchow and Wood 1980). Ludlow (1975) and Ng et al. (1975) concluded that water stress suspended the aging of physiologically young leaves and that upon rewatering, the photosynthetic rates of surviving leaves was greater than leaves of the same chronological age, but comparable to those of the same physiological age.

• Begg, J. and C. Buller 1995. An assessment of fiber hemp as an economic crop plant in Australia. Report for Co-operative Research Centre for Plant Science 11 pp.
• Begg, J. and N. C. Turner 1976. Crop water deficits. Advances in Agronomy 28: 161-217.
• Bridge, B. J. and M. J. Bell 1994. Effects of cropping on the physical fertility of krasnozems. in Proceedings of the Australian Institute of Agricultural Science Soils Ain’t Soils Conference Ulverstone, Tasmania, 5-6 September: 49-75.
• Chilvers, W. J. 1996. Managing Tasmania’s Cropping Soils: A practical guide for farmers. Department of Primary Industry and Fisheries, Tasmania, Australia: 80 pp.
• Dam, J. E. G. van 1995. Potentials of hemp as an industrial fibre crop. in Proceedings of First Bioresource Hemp Symposium, nova-Institut, Frankfurt, Germany, March 2-5: 405-422.
• Gates, C. T. 1955. The response of the young tomato plant to a brief period of water shortage 1. Whole plant and principle parts. Australian Journal of Biological Science 8: 196-214.
• Hackett, C. 1991. Mobilising environmental information about lesser known plants: the value of two neglected levels of description. Agroforestry Systems 14: 131-143.
• Leeper, G. W. and N. C. Uren 1993. Soil Science: An Introduction Melbourne University Press: 300 pp.
• Lisson, S.N. 1998. An integrated assessment of hemp (Cannabis sativa L.) and flax (Linum usitatissimum L.) as sources of fibre for newsprint production. Doctoral thesis, University of Tasmania: 211 pp.
• Muchow, R. C. and I. M. Wood 1980. Yield and growth responses of kenaf (Hibiscus cannabinus L.) in a semi-arid tropical environment to irrigation regimes based on leaf water potential. Irrigation Science 1: 209-222.
• Ludlow, M. M. 1975. Environmental and biological control of photosynthesis. (R. Marcelle, Ed.) Junk, The Hague: 123-134.
• Ng, T. T., J. R. Wilson and M. M. Ludlow 1975. The influence of water stress on water relations and growth of a tropical (C4) grass, Panicum maximum var. trichoglume. Australian Journal of Plant Physiology 2: 581-595.
• Salisbury, F. B. and C. W. Ross 1985. The photosynthesis-transpiration compromise. Plant Physiology Wadsworth Publishing Company: 540pp.