Study site.

The study was conducted in a mature planting at the North Willamette Research and Extension Center in Aurora, OR [lat. 45°16′47″N, long. 122°45′23″W; USDA plant hardiness zone 8b (U.S. Department of Interior, 2013)] from 2012 through 2014. The soil type at the site is Willamette silt loam (fine silty, mixed, superactive mesic Pachic Ultic Argixeroll). The field was certified organic by Oregon Tilth (Corvallis, OR), a USDA accredited agency, in 2012 (first fruiting year). See Harkins et al. (2013, 2014) for detailed information on site preparation and establishment and Dixon et al. (2015a, 2015b) for details on mature production.

Experimental design.

Treatments included cultivar (Marion and Black Diamond), irrigation (postharvest and no postharvest), weed management [nonweeded, hand-weeded, and weed mat (a porous, polyethylene groundcover)], and primocane training time (August and February). Treatments were arranged in a split-split-split plot design with five replicates. See Dixon et al. (2015b) for details of experimental plot layout. Plots were 1.5 × 3 m in size and contained four plants.

Weed management.

The three weed management strategies were applied to each plot individually. In nonweeded plots, weeds were allowed to grow after the first year (2010) and cut to soil level just before machine harvest (early July) during each harvest year (2012–14) to avoid any interference with the catcher plates on the machine harvester; the biomass removed was left in the row. In hand-weeded plots, weeds were removed as needed by hoeing throughout each growing season. The weed mat plots were covered with a 1.4-m-wide strip of black, woven, polyethylene groundcover (TenCate Protective Fabrics; OBC Northwest Inc., Canby, OR) centered on the row. Weeds were removed from the planting hole area and seams in the weed mat as required. More information on weed management strategies is provided in Dixon et al. (2015b).

Irrigation.

Each treatment was irrigated with a single lateral of drip tubing (UNIRAM; Netafim USA, Fresno, CA). The tubing had pressure-compensating emitters (1.9 L·h⁻¹ in-line) spaced every 0.6 m and was placed along the ground at the base of the plants under the weed mat, or was attached on a third wire on the steel posts, located 0.3 m above the ground in nonweeded and hand-weeded plots. The cultivar, irrigation, and weed management treatment combinations were irrigated independently using a manifold with electric solenoid valves and an automatic timer.

Irrigation was scheduled weekly based on estimates of crop evapotranspiration but was adjusted as needed each week to maintain similar leaf water potentials (LWP) among treatments (Dixon et al., 2015b). Irrigation was applied from 9 May to 8 Oct. 2012, 17 May to 27 Sept. 2013, and 28 May to 23 Sept. 2014 in the postharvest irrigation treatment. In the no postharvest irrigation treatment, irrigation was initiated on the same dates but withheld after the last fruit harvest date on 30 July 2012, 19 July 2013, and 15 July 2014. Thus, these plots received no effective water until the rainy season began on 12 Oct. 2012, 21 Sept. 2013, and 23 Sept. 2014.

Fertilization.

An Organic Materials Review Institute -approved fish hydrolysate and fish emulsion blend were diluted 1:3 (v/v) with water and applied through the drip system. In 2012, TRUE 402 (4N–0P–2K; True Organic Products, Inc., Spreckels, CA) was applied in four equal applications at a rate of 56 kg·ha⁻¹ N (Harkins et al., 2013). Converted Organics 421 (4N–2P–1K; Converted Organics of California LLC, Gonzales, CA) was used for the first four applications in 2013, and True Organics 512 (5N–0.4P–1.7K) was used for the last four applications in 2013 and all applications in 2014. The fertilizer(s) was split into eight equal applications (about every 2 weeks from 5 Apr. to 12 July 2013 and 19 Mar. to 25 June 2014) and applied at a total rate of 90 kg·ha⁻¹ N per year in 2013 and 2014 (based on the percentage of N listed on the label). Additional B fertilizer and lime and dolomite amendments were applied in 2013 and 2014 to correct nutritional deficiencies seen in soil tests and primocane leaf analyses (Dixon et al., 2015a). The total amount of nutrients applied to the planting are presented in Dixon et al. (2015a).

Primocane training.

Primocanes in the August-trained treatment were trained to the upper trellis wires on 13–14 Aug. 2012, 27–29 Aug. 2013, and 14 Aug. 2014, using the method described by Dixon et al. (2015b). In the February-trained treatment, primocanes were left on the wire for the drip irrigation lines, just aboveground level, throughout the growing season and the subsequent winter, until they were wrapped and tied to the upper two trellis wires on 21–25 Feb. 2013 and 21–28 Feb. 2014. Primocane training was done by replicate to avoid any possible date effects within treatment over the days required to train.

Data collection.

Primocanes (at 0.3 m height) were counted on two plants in each plot on 24 Jan. 2013, 20 Feb. 2014, and 18 Dec. 2014 (to assess primocane growth in 2012 to 2014, respectively). Individual primocanes were defined as originating at the crown or at a branch below 0.3 m and extending at least 1.0 m or to the first training wire.

One primocane was randomly cut from two plants per plot in Dec. 2014, weighed to determine the average individual fresh weight per cane in each plot, and then multiplied by cane number to estimate the total primocane fresh weight per plant in each plot. Subsamples that included tissue from the base, middle, and tip of the canes were analyzed for moisture content, C, N, phosphorus (P), K, calcium (Ca), Mg, S, iron (Fe), B, copper (Cu), manganese (Mn), zinc (Zn), and aluminum (Al) concentration (Dixon, 2015). A set of subsamples was also collected and analyzed for nutrients in Feb. 2014. Cane production was too low in 2014 to allow for destructive sampling, so cane fresh weight was not collected and subsamples were taken from near the end of the canes. The 2013 primocane nutrient concentrations were estimated as the average of the February and Dec. 2014 values. Percent moisture content was used to estimate dry biomass of the primocanes. The relationship between primocane number and dry biomass was determined in Dec. 2014: [primocane dry biomass = 0.13 × (no. of primocanes per plant) − 0.043; r² = 0.36; P < 0.0001] (Dixon, 2015). Primocane dry biomass was then calculated from the number of primocanes counted in Feb. 2013 and 2014. Primocane C and nutrient content were calculated from primocane nutrient concentrations and dry biomass in late winter. The 2013 primocane nutrient content refers to primocanes that grew in 2012 and then fruited as floricanes in 2013; 2014 primocane nutrient content refers to primocanes that grew in 2013 and then fruited as floricanes in 2014; and primocane nutrient content was calculated a final time in Dec. 2014 for primocanes that grew in 2014 and would have fruited as floricanes in 2015 (referred to as “2015 primocane nutrient content”).

Ripe fruit were harvested twice weekly from 24 June to 18 July in 2013 and 2014, using an over-the-row rotary harvester (Littau Harvesters Inc., Stayton, OR). Total yield was calculated from the weight of machine-harvested fruit on each date. A 25-berry subsample per treatment plot was shipped overnight to Brookside Laboratories (New Bremen, OH) on 8 July 2013 and 7 July 2014 and analyzed for C and nutrient concentrations and for percent moisture (Dixon et al., 2015a).

Senescing floricanes were removed by pruning at the base of the plant (≈0.1 m high) after fruit harvest on 29 July–5 Aug. 2013 and 30 July–1 Aug. 2014, per standard commercial practice (Strik and Finn, 2012). The total fresh biomass of the pruned floricanes was determined per plot, and a subsample of the pruned canes (comprised of cane, lateral, and node tissue from the lower, middle, and upper portion of the cane) was shipped overnight to Brookside Laboratories for analysis of total nutrient concentration and percent moisture content (Dixon et al., 2015a). Floricane dry biomass and C and nutrient content were then calculated. After pruning and data collection, the floricanes were left between the rows and flail-mowed (chopped), per standard commercial practice.

Dry biomass, C, and nutrient content data from the primocanes, fruit, and floricanes were used to calculate total aboveground plant nutrient content per hectare and accumulation and removal as affected by treatments in each year. Annual nutrient accumulation was calculated by subtracting the prior year’s primocane dry biomass (dormant weight) from the current year’s floricane dry biomass (to estimate floricane growth), then adding the dry biomass of the current year’s harvested fruit and new primocane growth. Annual nutrient removal was defined as the current year’s floricane prunings and fruit, while net change in aboveground nutrient content was defined as the nutrient accumulation minus the nutrient removal.

Data analysis.

Data were analyzed by year due to large differences in weather and winter damage observed in Dec. 2013 (Dixon et al., 2015b). Within year, data were analyzed for a split-split-split plot design with cultivar as the main plot factor, postharvest irrigation as the subplot factor, and weed management and training time as sub-subplots, using PROC MIXED in SAS (version 9.3; SAS Institute Inc., Cary, NC). Normality was assessed using a histogram of the residuals. Residuals were plotted to assess homogeneity of variance (residual by fitted value plot). Strong fanning or skewedness in the residual plots led to the data being log transformed for analysis to improve homogeneity of variance and to assess proportional effects. Data were back transformed for presentation. Means were compared for treatment effects using a Tukey’s honestly significant difference test at α = 0.05. Mean comparisons within significant interactions were done for treatments using least square means at α = 0.05.

Dry biomass production.

There were no main effects of cultivar or irrigation on the dry biomass of the dormant primocanes during the study. However, primocane dry biomass was significantly affected by an interaction between cultivar and irrigation in 2015, as well as by weed management each year, and by interactions between irrigation and training date in 2013 and cultivar and training date in 2015 (Tables 1–3). In 2013, the plants from the postharvest irrigation treatment had more dry biomass in the primocanes when the primocanes were trained in August than when they were trained in February. The plants also had more dry biomass in the primocanes with hand-weeding than with no weeding in 2 out of 3 years (2014 and 2015) and with weed mat each year. In 2015, ‘Marion’ had more primocane dry biomass than ‘Black Diamond’, particularly with postharvest irrigation, while ‘Black Diamond’ produced more primocane dry biomass when the primocanes were trained in February rather than in August.

Table 1.Dry biomass and total nutrient content in the dormant primocanes in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, 2013.^z

View Fullscreen

Table 2.Dry biomass and total nutrient content in the dormant primocanes in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, 2014.^z

View Fullscreen

Table 3.Dry biomass and total nutrient content in the dormant primocanes in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, 2015.^z

View Fullscreen

‘Marion’ produced more floricane dry biomass than ‘Black Diamond’ in 2013, regardless of weed management strategy (Table 4). In ‘Marion’, however, both weed control strategies increased floricane dry biomass compared with no weed control, whereas in ‘Black Diamond’, there was no difference between hand-weeded and no weed control. In 2014, floricane dry biomass was affected by all treatments (Table 5). Plants produced more floricane dry biomass when grown with postharvest irrigation than with no postharvest irrigation the previous year. ‘Marion’ continued to produce more floricane dry biomass than ‘Black Diamond’, but it was especially greater in ‘Marion’ when the primocanes were trained in February. In addition, training time only had an effect on floricane dry biomass in ‘Marion’. Weed control led to greater floricane dry biomass compared with no weeding when plants were trained in February. However, when plants were August-trained, there was no difference between hand-weeded and no weed control plots.

Table 4.Dry biomass and total nutrient content in the floricane prunings in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, July 2013.

View Fullscreen

Table 5.Dry biomass and total nutrient content in the floricanes prunings in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, July 2014.

View Fullscreen

In 2013, ‘Black Diamond’ produced more fruit dry biomass without than with postharvest irrigation the previous year, whereas the opposite was found in ‘Marion’ (Table 6). The plants also produced more fruit dry biomass with weed mat than with hand-weeding that year, and the least amount of fruit dry biomass without weed control. In 2014, ‘Marion’ produced more fruit dry biomass when the plants were trained in February rather than in August, while ‘Black Diamond’ produced the same amount of fruit dry biomass with either training treatment (Table 7). Both cultivars produced more fruit dry biomass with weed mat than with no weed control, although there was a larger difference between the weed mat treatment and the nonweeded treatment in ‘Black Diamond’ than ‘Marion’ (Table 7). Similarly to what was observed in 2013, ‘Black Diamond’ plants grown with weed mat produced the most fruit dry biomass, but August training resulted in a larger difference between weed control treatments. Fruit dry biomass was greater in weed mat than in hand-weeded plots only when plants were trained in August.

Table 6.Dry biomass and total nutrient content in the fruit of a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, 2013.

View Fullscreen

Table 7.Dry biomass and total nutrient content in the fruit of a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, 2014.

View Fullscreen

In 2013, the annual accumulation in total aboveground plant dry biomass (new primocane growth and floricane and fruit growth) was greatest, on average, in both cultivars with postharvest irrigation, weed mat, and February training (Table 8). In 2014, February training still led to the greatest accumulation in dry biomass, on average (Table 9). However, weed management only had an effect on dry biomass accumulation in August-trained plants, where plants grown with weed mat accumulated more than with other weed management strategies. The two cultivars accumulated similar amounts of dry biomass, averaging 4.1 and 4.7 t·ha⁻¹ in 2013 and 2014, respectively.

Table 8.Annual dry biomass and nutrient accumulation in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center, Aurora in OR, in 2013.

View Fullscreen

Table 9.Annual dry biomass and nutrient accumulation in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, in 2014.

View Fullscreen

In 2013, more dry biomass (harvested fruit and floricane prunings) was removed from ‘Marion’ than ‘Black Diamond’ when the plants were irrigated postharvest but not when there was no postharvest irrigation (Table 10). In addition, more dry biomass was removed from plants grown with weed mat, whereas the least was removed from those grown without weed control. In 2014, there was no effect of training time on dry biomass removal in ‘Black Diamond’, whereas more dry biomass was removed from ‘Marion’ when February-trained (Table 11). More dry biomass was also removed from ‘Marion’ than from ‘Black Diamond’ within each training time. Plants grown with weed mat had the greatest removal of dry biomass with August training; however, when plants were trained in February, there was no difference between dry biomass removal for weed mat and the hand-weeded treatment. The average removal in dry biomass was 3.5 t·ha⁻¹ in 2013 and 4.6 t·ha⁻¹ in 2014.

Table 10.Annual dry biomass and nutrient removal in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, in 2013.

View Fullscreen

Table 11.Annual dry biomass and nutrient removal in a mature planting of organic trailing blackberry located at the North Willamette Research and Extension Center in Aurora, OR, in 2014.

View Fullscreen

In 2013, net change in aboveground dry biomass ranged from 0.4 to 0.7 t·ha⁻¹ among the treatments, with the greatest increase observed in plants that were irrigated postharvest (P = 0.0156) and those that were February-trained (P = 0.0168; data not shown). In 2014, net change ranged from −0.3 to 0.3 t·ha⁻¹, although there were no significant treatment effects (data not shown).

Primocane nutrient content.

Treatment effects and interactions on dormant primocane macro- and micronutrient content varied among years (Tables 1–3). Cultivar and irrigation had limited direct effects on primocane nutrient content. In 2013, ‘Marion’ primocanes trained in February had less Cu in the tissue than those trained in August, or than ‘Black Diamond’ primocanes trained at either time (Table 1). In 2014, ‘Black Diamond’ primocanes contained less P and more S and Zn than ‘Marion’ primocanes (Table 2). In 2015, ‘Black Diamond’ primocanes had less K and B than ‘Marion’ primocanes trained in August, and less Mg, Fe, Cu, and Al than ‘Marion’ primocanes trained on either date (Table 3).

The impact of withholding irrigation after fruit harvest on primocane nutrient content often was affected by training time. In 2013, plants that received irrigation postharvest had less Ca and Zn in the primocanes trained in February as compared with August, whereas there was no effect of training time when the plants received no irrigation after fruit harvest (Table 1). In 2015, ‘Black Diamond’ plants that received postharvest irrigation had less S and Al in the primocanes than those that were not irrigated after harvest, whereas there was no difference in S or Al between the irrigation treatments in ‘Marion’.

In 2013, plants grown with weed mat had more primocane N, K (only when February-trained), Ca, S, and Cu than the nonweeded treatment. Plants in hand-weeded plots also contained more primocane Ca and Cu than the nonweeded treatment (Table 3). In 2014 and 2015, weed management affected all measured primocane nutrients. Both years, the plants with weed mat had a higher content of most nutrients in the primocanes than those in nonweeded plots, except for N and K in 2014 when the primocanes were August-trained, and a higher content of N and K (in February-trained only), P, and S than those in hand-weeded plots in 2014, but not in 2015 (Tables 4 and 5). In addition to the interactions mentioned previously, the 2013 primocanes that were trained in August had higher N, P, Ca, Mg, S, B, Cu, Mn, and Zn than when trained in February, while the opposite was found for Fe and Al content. In 2014, direct training effects were more limited, with only a similar response to training in the Fe and Al content. In addition, primocanes trained in February had a higher K content than those trained in August (Table 2). Only primocane Al was directly affected by training time in 2015, with a similar pattern as was seen in the previous 2 years.

Floricane nutrient content.

The nutrient content in the pruned floricanes was mostly affected by cultivar and weed management in 2013 (Table 4) and by cultivar, weed management, and training time in 2014 (Table 5). In both years, floricane nutrient content was higher in ‘Marion’ than ‘Black Diamond’ (except for P) and generally greater with weed control, although in 2014, plants in nonweeded and hand-weeded plots did not differ in floricane N and S (when August-trained), Zn (February-trained), Fe and Al (in ‘Black Diamond’), Cu (when irrigated postharvest), and P, Ca, Mg, and B. Use of weed mat led to higher floricane N, P, S, and Zn content than hand-weeding in 2013 and all nutrients in 2014, although only when August-trained for N, S, B, Zn, and Al, only when irrigated postharvest for Cu, and only in ‘Black Diamond’ for Fe.

The effect of training time, on average, was limited in 2013, with Mn content in floricane prunings only being higher when plants were trained in February (Table 4). However, in 2014, February training increased the content of all nutrients in the subsequent floricane prunings (Fe and Al only in hand-weeded plots) (Table 5). Floricane N, K, Ca, S, B, and Mn content were particularly high in ‘Marion’ for canes that were trained in February.

Fruit nutrient content.

Treatment effects on fruit nutrient content varied between 2013 and 2014 (Tables 6 and 7). In 2013, there was a direct effect of weed management on every nutrient. In general, plants growing in nonweeded plots had the lowest content of each nutrient in the harvested fruit, whereas plants in weed mat plots had the greatest, although hand-weeded and weed mat plots were not different for Ca, Mn, or Al. No other treatment had a direct effect on any nutrient in 2013 except plants that received postharvest irrigation had more fruit Al than those that did not. However, there was a cultivar × irrigation interaction on all fruit nutrients except N, Ca, Fe, and Cu. When plants received no postharvest irrigation, ‘Black Diamond’ fruit contained more P, K, Mg, S, B, Mn, and Al than did ‘Marion’. In plants irrigated postharvest, either the opposite was found (i.e., Mg, Mn, and Zn) or there was no difference between the cultivars (Table 6).

In 2014, there was no effect of irrigation on fruit nutrient content, but there was an effect of cultivar or a cultivar × training time interaction on the fruit content of each nutrient (Table 7). Fruit produced by August-trained ‘Marion’ plants had particularly low N, P, K, S, B, and Cu content than the other training time and cultivar combinations, whereas February-trained ‘Marion’ had a high P, Ca, Mg, Fe, Mn, and Zn content in the fruit. ‘Marion’ plants trained in February had a greater content of all nutrients, except for Al.

‘Black Diamond’ grown with weed mat in 2014 had a greater content of N, P, K, S, Zn, and Al in the fruit than in nonweeded plots and more content than hand-weeded plots in all nutrients except for Al. ‘Marion’ plants with weed mat only produced greater fruit N, P, K, and Zn than with no weed control. In both cultivars, use of weed mat increased fruit Ca, Mg, Fe, B, Cu, and Mn than no weed control and hand-weeded (except for Ca). February training increased fruit N, S, Fe, Mn, and Al content compared with August training. Fruit from plants in both training treatments had the highest P, K, Ca, Mg, B, Cu, and Zn when grown with weed mat, but August-trained plants tended to have a particularly low fruit nutrient content in the nonweeded treatment as compared with plants that were February-trained (with the exception of K and Mg). Plants in the hand-weeded treatment only had lower fruit P, K, Mg, B, Cu, and Zn content than the weed mat treatment when trained in August. In contrast, there was no difference in fruit nutrient content between the two methods of weed control when plants were trained in February.

A three-way interaction among irrigation, weed management, and training time revealed that fruit Ca content in 2014 was particularly low when the plants were not weeded, not irrigated after harvest, and trained in August when compared with those that were irrigated postharvest, trained in February, and had weed mat, or those that were not irrigated postharvest and were either trained in February and hand-weeded or trained in August and grown with weed mat (P = 0.0347; Fig. 1A). Fruit Fe content was particularly high in plants trained in February and either grown with weed mat and irrigated postharvest or hand-weeded and not irrigated postharvest when compared with plants that were not weeded and trained in August with either irrigation treatment or those that were hand-weeded and trained in August with postharvest irrigation (P = 0.0429; Fig. 1B).

View Figure

• Download as Powerpoint
• Download Figure

Nutrient accumulation and removal.

In 2013, the annual accumulation (new primocane and floricane growth and fruit harvest) in macronutrients per hectare was affected mostly by irrigation and weed management, whereas training time also affected micronutrient accumulation (Table 8). Postharvest irrigation increased the total accumulation of N in ‘Marion’ and of K, Mg, and S in both cultivars relative to no irrigation after harvest (Table 8). Weed control, particularly with weed mat, also increased the accumulation of many nutrients relative to no weeding, including N, P (weed mat only), K, Ca, Mg, S, Fe, B, Cu, Mn, Zn, and Al. Finally, February training increased the accumulation of Fe (particularly in plants that were irrigated postharvest), Cu, Mn, and Al relative to August training.

In 2014, ‘Marion’ plants accumulated more Mg, B, Cu, and Mn than ‘Black Diamond’ (Table 9). Plants grown with weed mat accumulated more Ca, Mg, Mn, and Zn than the other weed management strategies, regardless of training time. However, plants with weed mat only accumulated more N, P, K, S, and B than the other weed management treatments when August-trained; weed management did not affect nutrient accumulation in February-trained plants. Plants that were irrigated after harvest accumulated more B and Cu than those that were not irrigated.

More N, Ca, Mg, Fe, B, Cu, and Zn were removed from ‘Marion’ as a result of fruit harvest and floricane pruning than ‘Black Diamond’ in 2013 (Table 10). Removal of P and Al was also higher in ‘Marion’ plants irrigated after harvest than in ‘Black Diamond’ or in ‘Marion’ plants that were not irrigated after harvest. Nitrogen, P, K, S, Cu, and Zn removal was highest from plants grown with weed mat, whereas removal was lowest from those in nonweeded plots. In addition, either method of weed control increased the annual removal of Ca, Mg, Fe, B, Mn, and Al compared with nonweeded plants.

In 2014, removal was higher in ‘Marion’ plants that were trained in February for N, K, Ca, S, B, and Mn than in the other cultivar and training time combinations (Table 11). Phosphorus removal in ‘Marion’ was also higher when February-trained than when August-trained, although P removal was not higher than ‘Black Diamond’ for either training time. There was no effect of training time on nutrient removal in ‘Black Diamond’, which had less Mg, Fe, B, Zn, and Al (except in nonweeded plots) removed than ‘Marion’, irrespective of training time. Withholding irrigation after harvest decreased removal of N in both cultivars and of Cu in ‘Black Diamond’.

Less N and S was removed from ‘Black Diamond’ than ‘Marion’ when grown in nonweeded and hand-weeded plots, but there was no difference among cultivars when weed mat was used (Table 11). Plants grown with weed mat had a greater removal of Ca, Mg, and Mn than those in nonweeded and hand-weeded plots in both cultivars. Weed mat also led to greater removal of P, K, B, Zn, and Al when primocanes were trained in August, whereas there was generally no difference between weed mat and hand-weeded plots in the removal of these nutrients when primocanes were trained in February. In general, plants grown without weed control had the least nutrient removal.

In 2013, net change in aboveground nutrient content (annual accumulation minus annual removal) was 11–15 kg·ha⁻¹ N, 1–2 kg·ha⁻¹ P, 5–10 kg·ha⁻¹ K, 0–2 kg·ha⁻¹ Ca, 1–2 kg·ha⁻¹ Mg, 0–1 kg·ha⁻¹ S, 94–289 g·ha⁻¹ Fe, 6–11 g·ha⁻¹ B, 2–4 g·ha⁻¹ Cu, 33–68 g·ha⁻¹ Mn, 13–28 g·ha⁻¹ Zn, and 94–292 g·ha⁻¹ Al, depending on treatment (data not shown). Net nutrient content change was affected by irrigation and training time in 2013, but not by cultivar or weed management. Plots receiving postharvest irrigation accumulated more dry biomass and N, K, Ca, Mg, S, B, and Zn than those not receiving postharvest irrigation. February-trained plots accumulated more dry biomass, N, P, K, Ca, S, Fe, Cu, Mn, and Al than August-trained plots. There was a significant cultivar × irrigation effect on accumulation in aboveground Ca (P = 0.0428). All ‘Marion’ plots and postharvest-irrigated ‘Black Diamond’ plots accumulated an average of 1.7 kg·ha⁻¹ of Ca in 2013, whereas 0.4 kg·ha⁻¹ Ca was removed from ‘Black Diamond’ plots without postharvest irrigation.

There were few treatment effects on change in aboveground nutrients in 2014 (data not shown). The net aboveground change of K and S was greater in ‘Black Diamond’ (−8 and −0.6 kg·ha⁻¹, respective removal) than ‘Marion’ (−3 and −0.2 kg·ha⁻¹, respective removal). In addition, there was a net accumulation of Cu (8 g·ha⁻¹) in ‘Marion’, whereas there was a removal in ‘Black Diamond’ (−4 g·ha⁻¹). More Fe and Al were removed from plants that were trained in February (−390 and −389 g·ha⁻¹, respectively) than those trained in August (−112 and −114 g·ha⁻¹, respectively). Aboveground net nutrient change ranged from −8 to −13 kg·ha⁻¹ N, 0 to −1 kg·ha⁻¹ P, −3 to −8 kg·ha⁻¹ K, 0 to 2 kg·ha⁻¹ Ca, 0 to −1 kg·ha⁻¹ Mg, 0 to −1 kg·ha⁻¹ S, −112 to −390 g·ha⁻¹ Fe, −9 to 4 g·ha⁻¹ B, −4 to 8 g·ha⁻¹ Cu, −38 to 8 g·ha⁻¹ Mn, −25 to 10 g·ha⁻¹ Zn, and −114 to −478 g·ha⁻¹ Al, depending on treatment.

Carbon.

Carbon content of the primocanes, floricanes, and fruit varied between years and followed the same pattern as dry biomass in almost every case, so specific results are not presented in detail here (Tables 1–5). Aboveground accumulation and removal in C content was similar in 2013 (Tables 8 and 10) and 2014 (Tables 9 and 11). Treatment effects differed from dry biomass in two cases. In 2013, irrigation only affected ‘Marion’ plants, which accumulated more C with postharvest irrigation than without, and in 2014, ‘Marion’ accumulated more C than ‘Black Diamond’, regardless of irrigation strategy.

In 2013, irrigation and training time had direct effects on net change in aboveground C (data not shown). Plots that received irrigation after harvest accumulated 0.4 t·ha⁻¹ C compared with 0.2 t·ha⁻¹ C when not irrigated postharvest. The net accumulation of February-trained plots was 0.4 t·ha⁻¹ C, whereas August-trained plots accumulated 0.2 t·ha⁻¹ C. There were no treatment effects on net aboveground C in 2014, with an average measured net removal of 0.03 t·ha⁻¹ of C (data not shown).

The aboveground C stock of the treatments averaged 0.4–1.1 t·ha⁻¹ during dormancy in late winter. The greatest C stock was measured in ‘Black Diamond’ grown without postharvest irrigation, with weed mat, and when trained in February, whereas the lowest was measured in ‘Black Diamond’ grown without postharvest irrigation, without weed control, and when trained in August. In ‘Marion’, the highest aboveground C stock occurred when plants received no irrigation after harvest, were grown with weed mat, and were August-trained, whereas the lowest occurred with postharvest irrigation, no weed control, and February training.