4. Discussion
4.1. Effectsoffinerootdiameter,samplingdepthandinclusionof the understorey vegetation on FRB estimates
Despite the fact that species differ in the tissue structure and morphology of their roots, there was a clear linear relationship between the root diameter classes, and the relationships did not differ between the deciduous and needleleaf forests (Fig. 1). The close relationship between diameter classes could be a result of the autocorrelation between diameter classes, i.e. ≤1 mm was included in the ≤2 mm diameter class and ≤2 mm in the ≤5 mm class. The regression equations had a somewhat smaller error between the ≤1 mm and ≤2 mm diameter classes than between the ≤2 mm and ≤5mm diameter classes (Fig. 1). Our results are not fully com- parable to the earlier ones. Leuschner and Hertel (2003) found significant differences between tree species in the contribution of the 2–5 mm diameter class to the ≤5 mm diameter FRB when they compared only three tree species, and Chen et al. (2004) showed strong linear relationships between ≤2 mm and ≤5 mm FRB when they made the comparisons on tree level for different tree species. Our dataset was larger than the earlier ones and it covered data on several tree species at stand level.
The equations allowed us to standardize all the FRB data in our database to the ≤2 mm diameter class, and to render the results published for the ≤1 mm, ≤2 mm and ≤5 mm diameter classes com- parable (Table 3). Attempts have been made to introduce a common definition for fine roots (Böhm, 1979; Zobel and Waisel, 2009), but there is no consensus, and many recent studies have focused more on ≤0.5 mm roots because they are physiologically more active than larger ones (Pregitzer et al., 1998; Joslin et al., 2006; Yanai et al., 2006; Park et al., 2008; Makita et al., 2009).
The various fine root fractions were treated together in many of the earlier reviews, and this undoubtedly increased the varia- tion and resulted in biased estimates for FRB (Leuschner and Hertel, 2003). In our database, however, the standardized ≤2 mm FRB esti- mates did not differ significantly from those calculated using the original data, which contained diameter classes ≤1 mm, ≤2 mm and ≤5 mm (p > 0.05) (Table 3). This was probably due to the fact that most of the data sets we used included FRB data for the ≤2 mm diameter class, and the effect of the ≤1 mm FRB data could be off- set by that of the ≤5 mm FRB data. The total FRB estimates would have been 25–29% smaller if we had standardized the results to the ≤1 mm diameter class, and 34–60% higher in the case of the ≤5 mm diameter class (Table 3).
We were able to find data to produce estimates for the mean understorey vegetation FRB for needleleaf and deciduous forests in both the temperate and boreal zones. Since the FRB production and turnover rates of different species groups vary (Finér et al., 1997; Finér and Laine, 1998; Majdi and Andersson, 2005), it is important to determine FRB separately for the trees and the understorey vege- tation, as the roots of the latter contributed 10–30% of the total FRB (Table 5), this proportion being highest in the boreal and needleleaf forests. The observation that there were relatively more temper- ate broadleaf than needleleaf forests without any understorey FRB indicates that the FRB sampling was carried out in dense temperate forest stands where there is not enough light for any understorey vegetation to grow. Sampling in spring, before the full development of the leaves, could probably have given different estimates. We did not find any studies from tropical forests in which the FRB of trees had been determined separately from that of the understorey veg-
etation, but the diversity of species and canopy layers are much higher in pristine tropical forests than that in temperate or boreal forests or the understorey can mostly consist of juvenile trees, which makes separation too laborious and somewhat arbitrary. There was large variation in the proportion of total FRB attributable to trees, and in contrast to the results of Chen et al. (2004), we did not find any correlation between this proportion and tree stand characteristics. Chen et al. (2004) also studied the relationship between the proportion of FRB attributable to trees in boreal and temperate forests, and one reason why they found a correlation and we did not may be that they made separate calculations for shade-tolerant and shade-intolerant species, which we could not do. Studies on the dynamics of above-ground biomass have shown clear relationships between the understorey vegetation biomass and tree stand characteristics such as basal area, stand age and stand density, and these relationships have varied between under- storey species groups (Alaback, 1982; Muukkonen and Mäkipää, 2006). Such species-specific relationships may counteract the rela- tionship between the proportion of total FRB accounted for by trees and the FRB for the understorey vegetation.
According to the review by Schenk and Jackson (2002), rooting depths of 58 cm in boreal forests, 104–121 cm in temperate forests and 91–94 cm in tropical forests include 95% of FRB on average, the corresponding depths for 50% of FRB being 12 cm, 21–23 cm and 14–16 cm, respectively. This suggests that the studies included in our database covered 50% of the FRB in most cases and seldom the whole rooting depth (Table 4). The sampling depth was deeper in the temperate and tropical forests than in the boreal forests (Table 4), which could indicate that the rooting depth in the boreal forests was not as deep as in the other biomes. These rooting depths are generalizations over wide geographical scales and there is much variation within each biome caused by soil properties like soil type, stoniness and impermeable layers (e.g. Schenk and Jackson, 2002) and species (Stone and Kalisz, 1991), which we could not take into account in our analyses.
Our FRB estimates based on the original sampling depths fell within the same range as reported in the earlier studies in the case of the temperate and tropical forests but were higher for the boreal forests (Tables 6 and 8), whereas our estimates extrapolated for the whole rooting depth were in general higher than those reported earlier, even including the reports that also covered deeper soil layers (Jackson et al., 1996, 1997; Leuschner and Hertel, 2003). Our results, like the earlier ones based on the original sampling depths, are most probably underestimates of the true FRB. The FRB esti- mates extrapolated with the equations and parameters presented by Gale and Grigal (1987) and Jackson et al. (1997) are higher than those based on the original sampling depths, but then Jackson et al. (1997) produced their parameters with only a small body of data and without separating the tree roots from the understory veg- etation roots. We introduced error into our estimates when we applied the equations and parameters, and more studies on FRB covering the whole rooting depth in different forest biomes would be needed to estimate the magnitude of this error, so that the present results must be interpreted with caution. Most of the ear- lier studies nevertheless indicate that the FRB is higher in temperate broadleaf forests than in temperate needleleaf forests, and higher in evergreen tropical forests than in deciduous ones (Table 7). In our analyses the FRB of trees based on the original rooting depth was significantly higher in the temperate broadleaf forests than in the temperate needleleaf forests and the total FRB extrapolated to the whole rooting depth was significantly smaller in the boreal zone than in the other zones, the results depending greatly on the rooting depth used for calculating the FRB. The results do not fully confirm that there are differences in FRB between the biomes and more data covering the whole rooting depth are needed to find that out.
4. สนทนา4.1. Effectsoffinerootdiameter, samplingdepthandinclusionof understorey พรรณนาใน FRB ประเมินทั้ง ๆ ที่พันธุ์แตกต่างกันในโครงสร้างเนื้อเยื่อและสัณฐานวิทยาของรากของตน มีความสัมพันธ์เชิงเส้นชัดเจนระหว่างชั้นเส้นผ่าศูนย์กลางราก และความสัมพันธ์ที่ได้ไม่แตกต่างกันระหว่างป่าไม้ผลัดใบและ needleleaf (Fig. 1) ความสัมพันธ์ใกล้ชิดระหว่างชั้นเส้นผ่าศูนย์กลางอาจเป็นผลลัพธ์ของ autocorrelation ระหว่างชั้นเส้นผ่าศูนย์กลาง เช่นรวมอยู่ ≤1 มม. ≤2 มม.เส้นผ่าศูนย์กลางคลาสและคลา ≤5 มม. ≤2 มม. สมการถดถอยได้มีข้อผิดพลาดค่อนข้างเล็กระหว่าง ≤1 mm และ ≤2 mm เส้นผ่าศูนย์กลางคลากว่าระหว่าง ≤2 mm และ ≤5mm เส้นผ่าศูนย์กลางคลา (Fig. 1) ผลของเราทั้งหมดไม่ใช่ com-นิทานคติสอนใจการก่อนหน้านี้ Leuschner และ Hertel (2003) พบสำคัญความแตกต่างระหว่างแผนภูมิชนิดในสัดส่วนของคลาเส้นผ่าศูนย์กลาง 2-5 มม.ใบ ≤5 มม.เส้นผ่าศูนย์กลาง FRB เมื่อพวกเขาเปรียบเทียบเพียงสามพันธุ์ต้นไม้ Chen et al. (2004) แสดงให้เห็นว่าความสัมพันธ์เชิงเส้นแข็งแรงระหว่าง ≤2 mm และ ≤5 mm FRB เมื่อทำการเปรียบเทียบระดับภูมิสำหรับชนิดแผนภูมิอื่น ชุดข้อมูลของเรามีมากกว่าคนก่อนหน้านี้ และจะครอบคลุมข้อมูลในชนิดแผนภูมิหลายระดับขาตั้งThe equations allowed us to standardize all the FRB data in our database to the ≤2 mm diameter class, and to render the results published for the ≤1 mm, ≤2 mm and ≤5 mm diameter classes com- parable (Table 3). Attempts have been made to introduce a common definition for fine roots (Böhm, 1979; Zobel and Waisel, 2009), but there is no consensus, and many recent studies have focused more on ≤0.5 mm roots because they are physiologically more active than larger ones (Pregitzer et al., 1998; Joslin et al., 2006; Yanai et al., 2006; Park et al., 2008; Makita et al., 2009).The various fine root fractions were treated together in many of the earlier reviews, and this undoubtedly increased the varia- tion and resulted in biased estimates for FRB (Leuschner and Hertel, 2003). In our database, however, the standardized ≤2 mm FRB esti- mates did not differ significantly from those calculated using the original data, which contained diameter classes ≤1 mm, ≤2 mm and ≤5 mm (p > 0.05) (Table 3). This was probably due to the fact that most of the data sets we used included FRB data for the ≤2 mm diameter class, and the effect of the ≤1 mm FRB data could be off- set by that of the ≤5 mm FRB data. The total FRB estimates would have been 25–29% smaller if we had standardized the results to the ≤1 mm diameter class, and 34–60% higher in the case of the ≤5 mm diameter class (Table 3).We were able to find data to produce estimates for the mean understorey vegetation FRB for needleleaf and deciduous forests in both the temperate and boreal zones. Since the FRB production and turnover rates of different species groups vary (Finér et al., 1997; Finér and Laine, 1998; Majdi and Andersson, 2005), it is important to determine FRB separately for the trees and the understorey vege- tation, as the roots of the latter contributed 10–30% of the total FRB (Table 5), this proportion being highest in the boreal and needleleaf forests. The observation that there were relatively more temper- ate broadleaf than needleleaf forests without any understorey FRB indicates that the FRB sampling was carried out in dense temperate forest stands where there is not enough light for any understorey vegetation to grow. Sampling in spring, before the full development of the leaves, could probably have given different estimates. We did not find any studies from tropical forests in which the FRB of trees had been determined separately from that of the understorey veg-etation, but the diversity of species and canopy layers are much higher in pristine tropical forests than that in temperate or boreal forests or the understorey can mostly consist of juvenile trees, which makes separation too laborious and somewhat arbitrary. There was large variation in the proportion of total FRB attributable to trees, and in contrast to the results of Chen et al. (2004), we did not find any correlation between this proportion and tree stand characteristics. Chen et al. (2004) also studied the relationship between the proportion of FRB attributable to trees in boreal and temperate forests, and one reason why they found a correlation and we did not may be that they made separate calculations for shade-tolerant and shade-intolerant species, which we could not do. Studies on the dynamics of above-ground biomass have shown clear relationships between the understorey vegetation biomass and tree stand characteristics such as basal area, stand age and stand density, and these relationships have varied between under- storey species groups (Alaback, 1982; Muukkonen and Mäkipää, 2006). Such species-specific relationships may counteract the rela- tionship between the proportion of total FRB accounted for by trees and the FRB for the understorey vegetation.ตามตรวจทานโดย Schenk และ Jackson (2002), rooting ลึก 58 ซม. boreal ป่า 104 – 121 ซม.ในป่าแจ่มและ 91-94 ซม.ในป่าเขตร้อนรวม 95% FRB โดยเฉลี่ย ความลึกที่สอดคล้องกันสำหรับ 50% ของ FRB ถูก 12 ซม. 21 – 23 เซนติเมตร และ 14-16 ซม. ตามลำดับ แนะนำให้ ศึกษาข้อมูลของเราครอบคลุม 50% ของ FRB ในกรณีส่วนใหญ่และแทบทั้งหมด rooting ลึก (ตาราง 4) สุ่มตัวอย่างความลึกลึกในป่าเขตร้อน และแจ่มกว่าในป่า boreal (ตาราง 4), ซึ่งอาจบ่งชี้ว่า ความลึกในป่า boreal rooting ไม่เป็นลึกใน biomes อื่น ๆ Rooting ลึกได้แก่ generalizations เหนือระดับทางภูมิศาสตร์กว้าง และมีความผันแปรมากในแต่ละชีวนิเวศเกิดจากคุณสมบัติของดินเช่นดินชนิด stoniness และการซึมผ่านของชั้น (เช่น Schenk และ Jackson, 2002) และสปีชีส์ (หินและ Kalisz, 1991), ซึ่งเราไม่สามารถนำเข้าบัญชีในการวิเคราะห์ของเราOur FRB estimates based on the original sampling depths fell within the same range as reported in the earlier studies in the case of the temperate and tropical forests but were higher for the boreal forests (Tables 6 and 8), whereas our estimates extrapolated for the whole rooting depth were in general higher than those reported earlier, even including the reports that also covered deeper soil layers (Jackson et al., 1996, 1997; Leuschner and Hertel, 2003). Our results, like the earlier ones based on the original sampling depths, are most probably underestimates of the true FRB. The FRB esti- mates extrapolated with the equations and parameters presented by Gale and Grigal (1987) and Jackson et al. (1997) are higher than those based on the original sampling depths, but then Jackson et al. (1997) produced their parameters with only a small body of data and without separating the tree roots from the understory veg- etation roots. We introduced error into our estimates when we applied the equations and parameters, and more studies on FRB covering the whole rooting depth in different forest biomes would be needed to estimate the magnitude of this error, so that the present results must be interpreted with caution. Most of the ear- lier studies nevertheless indicate that the FRB is higher in temperate broadleaf forests than in temperate needleleaf forests, and higher in evergreen tropical forests than in deciduous ones (Table 7). In our analyses the FRB of trees based on the original rooting depth was significantly higher in the temperate broadleaf forests than in the temperate needleleaf forests and the total FRB extrapolated to the whole rooting depth was significantly smaller in the boreal zone than in the other zones, the results depending greatly on the rooting depth used for calculating the FRB. The results do not fully confirm that there are differences in FRB between the biomes and more data covering the whole rooting depth are needed to find that out.
การแปล กรุณารอสักครู่..
