Methodological Comparisons of Absorptive and Transport Fine Root Production, Mortality and Decomposition in A Loblolly Pine Plantation Forest

10 Background and aims Fine roots can be functionally classified into an absorptive fine root pool (AFR) and a 11 transport fine root pool (TFR) and their production, mortality and decomposition play a critical role in forest soil 12 carbon (C) cycling. Different methods give significant estimates. However, how methodological difference affects 13 AFT and TFR production, mortality, and decomposition estimates remains unclear, impeding us to accurately 14 construct soil C budgets. 15 Methods We used dynamic-flow model, a model combining measurements of litterbags and soil cores, and 16 balanced-hybrid model, a model combining measurements of minirhizotrons and soil cores, to quantify these fine 17 root estimates in a managed loblolly pine forest. 18 Results Temporal changes in production, mortality or decomposition estimates using both models were not different 19 for both AFRs and TFRs. Annual production, mortality, and decomposition were comparable between AFRs and 20 TFRs when measured using the dynamic-flow model but significantly higher for AFRs than for TFRs when 21 measured using the balanced-hybrid model. Annual production, mortality and decomposition estimates using the 22 balanced-hybrid model were 75%, 71% and 69% higher than those using the dynamic-flow model ( P < 0.05 for all), 23 respectively, for AFRs, but 12%, 6% and 5% higher than those using the dynamic-flow model ( P > 0.05 for all), dynamic-flow model. Lower AFR estimates using the dynamic-flow model appeared to result from the 26 underestimated AFR mass loss rate induced by the litterbag method. 27 Conclusions Methodological difference had a more significant impact on AFR estimates than on TFR estimates. 28 These results have important implications for better quantifying the most dynamic fraction of fine root system and 29 understanding soil C cycling. 30 The study was conducted in a commercially managed loblolly pine ( P. taeda L.) forest (35º48'N 76º40'W) located in the lower coastal plain of Washington County, North Carolina, USA. Mean annual precipitation and temperature for 79 the period 2011-2017 were 1320mm and 12.2 °C, respectively. The topography of the area is flat (<5m above sea 80 level) and on a Belhaven series histosol soil (loamy mixed dysic thermic terric Haplosaprists). The study area was 81 harvested of trees and ditched/drained in the late 19th to the early 20th century before being converted to a 82 commercial pine plantation. The forest was fertilized with nitrogen and phosphorus at the time of planting and mid- 83 rotation. The soil C and nitrogen concentrations at 20cm depth were 26% and 1.0%, respectively. The mean canopy 84 height, diameter at the breast height, and stand age during the study period were approximately 24 m, 33cm, and 23 85 years, respectively. For a full site description, refer to Noormets et al. (2010). Three plots, about 5m ×9m for each 86 and 100m to 800m apart, were established at random in the plantation in 2013. Only loblolly pine fine roots were 87 studied as they accounted for over 90% of total fine root mass in this forest.


39
The conventional ingrowth core and soil core methods, which are low cost and ready-to-use, had been 40 extensively applied to assess fine root production and mortality (Vogt et al. 1998

54
Fine roots have been traditionally defined as distal roots with diameters <2mm. Recent studies have shown that 55 the hierarchical root system is morphologically, chemically and functionally heterogeneous and can be partitioned  ). An improved understanding of AFR and TFR dynamics in loblolly pine plantation forests is critical for 68 developing silvicultural and rotation strategies to increase C sequestration capacity.

69
In this study, we used the soil core method, litterbags, and minirhizotrons to assess biomass and necromass 70 dynamics, mass loss pattern and growth and death rates of AFRs and TFRs in a managed loblolly pine forest. The 71 objectives were to 1) use both the dynamic-flow model and the balanced-hybrid model to quantify AFR and TFR 72 production, mortality, and decomposition in this forest, 2) assess to what extent methodological difference affects 73 AFR and TFR estimates and 3) determine which method is more reliable.

75
The study was conducted in a commercially managed loblolly pine (P. taeda L.) forest (35º48'N 76º40'W) located in 78 the lower coastal plain of Washington County, North Carolina, USA. Mean annual precipitation and temperature for 79 the period 2011-2017 were 1320mm and 12.2 °C, respectively. The topography of the area is flat (<5m above sea 80 level) and on a Belhaven series histosol soil (loamy mixed dysic thermic terric Haplosaprists). The study area was 81 harvested of trees and ditched/drained in the late 19th to the early 20th century before being converted to a 82 commercial pine plantation. The forest was fertilized with nitrogen and phosphorus at the time of planting and mid-83 rotation. The soil C and nitrogen concentrations at 20cm depth were 26% and 1.0%, respectively. The mean canopy 84 height, diameter at the breast height, and stand age during the study period were approximately 24 m, 33cm, and 23 85 years, respectively. For a full site description, refer to Noormets et al. (2010). Three plots, about 5m ×9m for each 86 and 100m to 800m apart, were established at random in the plantation in 2013. Only loblolly pine fine roots were 87 studied as they accounted for over 90% of total fine root mass in this forest. . Previous study showed that over 94 90% of fine roots were distributed in 0 -30 cm soil layer. Collected soil cores were rinsed with clean tap water 95 through a 0.5mm mesh sieve to isolate roots. We only studied loblolly pine fine roots as they accounted for over 96 95% of total fine root mass. Loblolly pine fine roots with light color and intact stele and periderm were regarded as 97 live roots, while those with dark color and damaged stele and periderm were dead ones. In this study, AFRs 98 represented the first and second-order roots, while TFRs were third-order roots and higher with diameter <2mm.

118
Fine root mass loss pattern was simulated using an exponential equation with only two parameters: where y(t) and y0 are root mass at time t (year) and the start, respectively. The two parameters λ (year -1 ) and k

121
(year -1 ) were calculated based on the fine root mass remaining in litterbags collected on all sampling occasions using 122 nonlinear regression. e −k t is fine root decomposition rate which is time-dependent. It is the highest at the beginning 123 and decreases over time.

124
The fine root mortality rate in interval i (μi) was assumed to be constant. The total production (gi) mortality (mi) 125 and decomposition (di) in interval i were calculated by the following equations, where Bi (0) was an exponential integral function (Abramowitz and Stegun, 1964, ch. 6).

170
Model test

171
The efficacy of the models for estimating the production, mortality, and decomposition was tested by comparing the 172 predicted with the measured AFR and TFR biomass in July using a subset of data not used for model accuracy. The predicted AFR and TFR biomass in July were calculated according to the procedures described in 175 Hendrick and Pregitzer (1993) and Hendricks et al. (2006).

177
Statistical analysis

178
The plots were considered as replicates (n = 3), and data collected (sub-replicates) within the same plot were 179 averaged before performing statistical analysis. One-way ANOVA or paired t-test was used to assess the differences

191
Mass loss rate

192
Live AFR substrates had significantly higher percent mass remaining than live TFR substrates at the late 193 decomposing stage, but dead AFR and TFR substrates had comparable percent mass remaining during the whole 194 study period (Fig. 2). All live root substrates decomposed significantly faster than dead root substrates (Fig. 2).

195
Temporal changes in fine root production, mortality and decomposition rates were generally the same between the 198 two models, with greater production in warmer months and greater mortality and decomposition occurring in cooler 199 months (Fig. 3). Production, mortality, and decomposition rates using dynamic-flow model were comparable 200 between AFRs and TFRs at all intervals. In contrast, production, mortality, and decomposition rates using the 201 balanced-hybrid model were significantly higher for AFRs than for TFRs in most intervals (Fig. 3). AFR production, 202 mortality and decomposition rates using the dynamic-flow model were significantly lower than those using the 203 balanced-hybrid model in most intervals, while TFR production, mortality and decomposition rates were not 204 significantly different between the two models in all intervals (Fig. 3).

206
Annual fine root estimates

207
Annual production, mortality, and decomposition were comparable between AFRs and TFRs in dynamic-flow 208 model estimation but significantly higher for AFRs than for TFRs in balanced-hybrid model estimation (Fig. 4).

217
Model Test

218
The measured AFR biomass in July was 28% and 15% higher than that estimated by the dynamic-flow model and 219 the balanced-hybrid model, respectively, while the measured TFR biomass in July was 19% and 11% higher than

226
2020b). Failing to assess the biomass and necromass dynamics impedes us to characterize soil C flux dynamics 227 through AFR and TFR growth, death and decay. In this managed loblolly pine forest, AFRs had significantly lower 228 biomass than TFRs but made comparable or even significantly greater contributions to total fine root production, 229 mortality and decomposition than TFRs did, demonstrating that three-dimensional, function-based study is essential 230 to accurately quantify fine root C budget.

231
Different methods yielded divergent fine root estimates, but all these methodological comparisons were . This knowledge gap has hindered us to better identify the strengths and weaknesses of each method and 234 characterize the C allocation pattern within the root system. Our study for the first time used two types of models, a 235 litterbag-based model and a minirhizotron-based model, to assess AFR and TFR production, mortality, and 236 decomposition. AFR estimates were significantly more responsive to methodological difference than TFR estimates.

237
Thus, methodological difference impact must be taken into account when assessing AFR and TFR dynamics.

238
Model test showed that the balanced-hybrid model had greater estimation accuracy than the dynamic-flow 239 model. This can be explained by the inherent differences between them. In the balanced-hybrid model, the relative 240 production and mortality rates at the tube-soil interface are assumed to be representative of those in bulk soil. This  The balanced-hybrid has a greater estimation accuracy than the dynamic-flow model and differences between the 273 two models did not significantly affect TFR estimates but significantly affected AFR estimates. Thus, the 274 methodological difference must be considered to accurately characterize AFR and TFR dynamics and to quantify

383
Note: AFR production, mortality and decomposition estimates using balanced-hybrid model have been reported. We  Absorptive (AFR) and transport (TFR) ne root biomass and necromass dynamics (g m-2 for the 0-0.30 m soil depth; n= 3; mean ± SE). Note: AFR biomass and necromass have been reported. We use these values for the purpose of comparison.

Figure 2
Mass loss patterns of live and dead absorptive (AFR) and transport (TFR) ne root substrates measured using litterbags in a managed loblolly pine forest (n= 3; mean ± SE; different letters stand for signi cant difference in means, P<0.05).

Figure 3
Temporal changes in production, mortality and decomposition estimates of absorptive (AFR) and transport (TFR) ne roots using balanced-hybrid model (BH) and dynamic-ow model (DF) in a managed loblolly pine plantation forest (n= 3; mean ± SE). Different letters stand for signi cant difference in means (P<0.05). Note: AFR production, mortality and decomposition estimates using balanced-hybrid model have been reported. We use these values for the purpose of comparison. Annual absorptive (AFR) and transport (TFR) ne root production, mortality and decomposition measured using balanced-hybrid model (BH) and dynamic-ow model (DF) in a managed loblolly pine plantation forest (n= 3; mean ± SE). Different letters stand for signi cant difference in means (P<0.05).