The research site of the Kiel Ecosystem Research Center (Hörmann et al., 1992) is located in the Bornhöved Lake District, 30 km south of Kiel (54°06′ N, 10°14′ E). The alder forest, growing on the west shore of Lake Belau is an Alnetum glutinosae. In 1992 the trees of the edge zone under investigation had an average tree age of 45 y. The forest is typical of the North German landscape, representing conditions highly influenced by near-surface groundwater dynamics. The soil properties are summarized in Table 1.
Field measurements of total soil respiration and rhizomicrobial respiration
Soil CO2 efflux can be measured by several techniques, reviewed and compared by Janssens et al. (2000). In the present study an open dynamic system was used. The system (Fig. 1) comprised eight parallel channels, each with its own measuring and reference gas units consisting of a pump (WISA, Wuppertal, Germany), a mechanical flow controller (Krohne, Düsseldorf, Germany), and a magnetic valve (Herion, Fellbach, Germany). This technical device allowed both overpressure and underpressure to be applied to the chambers. Behind the magnetic valve the air stream was passed through an electronic flowmeter (Tylan General, USA) and a gas cooling unit (Walz, Effeltrich, Germany) to an infrared gas analyzer (Fisher-Rosemount, Hanau, Germany).
Figure 1. System developed for field measurements of total soil respiration and rhizomicrobial respiration. Abbreviations: P, pump; mFC, mechanical flow controller; mV, magnetic valve; eFM, electronic mass flowmeter; C, gas cooler (for drying the air); IRGA, infrared gas analyzer.
Download figure to PowerPoint
Ambient air was continuously sucked through the chamber with a flow rate of 1 l min−1. The diameter of the inlet was 3 cm, which was large enough to avoid an underpressure higher than 1 µbar in the system (pressure sensor type 223, MKS Baratron, München, Germany, for details see: Kutsch, 1996). The channels were measured subsequently at 3 min intervals. In parallel, ambient air was passed through a reference tube to the gas analyzer measuring in the differential mode. The chambers having about 2.8 l volume covered an area of 16 × 12.5 cm. To measure CO2 flux they were attached to aluminum frames which were fixed in the soil. Each measurement period lasted no longer than 3 d to minimize chamber-induced changes in soil moisture. After this period the chamber was attached to another frame. The frames remained at the same place during the whole growing season of 1992.
We preferred an open dynamic system in this study because it enabled continuous measurements as close as possible to the natural conditions. In open systems: CO2-concentration in the chamber is close to ambient; pressure difference between inside and outside the chamber can be minimized; and pressure fluctuations induced by the vertical wind component are transmitted to the soil surface. The latter induce an increase of the mass flow of soil air through the soil surface and can increase the efflux of CO2 during periods of high wind speed (Janssens et al., 2000; Rayment & Jarvis, 2000). However, the evaluation of the data showed that our system not only transmitted but also modified the natural pressure fluctuations caused by an inappropriate design of the chamber inlet (Kutsch, 1996). Placing the inlet at one side of the chamber caused an overpressure inside the chamber when the wind blew directly onto the inlet and an underpressure inside the chamber in all other directions. Fig. 2 shows this dependency of chamber internal pressure on wind speed and wind direction measured in the laboratory using a fan. Two consequences were drawn for the data evaluation: data measured when wind was directed to the inlet were rejected; and measurements made during periods with wind speed > 3.5 m s−1 causing an underpressure lower than −20 µbar were interpreted as ‘influenced more by a chamber effect than by natural pressure fluctuations’ and were also eliminated. In late 1992 these problems were avoided by covering the inlets of the chambers by a dome-shaped, gauze-covered shelter (Fig. 3). In addition, a second chamber was inserted in the reference gas stream. This closed chamber adjusted the volume of the reference gas stream to that of the measuring gas stream in order to prevent fluctuations in the flux signal after rapid changes in the CO2-concentration of the ambient air (Kutsch, 1996). The new chamber design showed no dependency of the pressure fluctuations inside the chamber on wind direction during laboratory tests.
Figure 2. Pressure difference between chamber and ambient air in relation to wind speed and wind direction. When wind comes from the direction of the inlet it caused an overpressure up to more than 100 µbar inside the chamber at 5-m s −1 . Other directions caused underpressure down to −30 µbar at 5 m s −1 . Measurements were taken in the laboratory using a fan.
Download figure to PowerPoint
Figure 3. Modified chamber design used after autumn 1992. The inlet of the chamber is sheltered by a gauze-covered dome. The additional chamber adjusts the volume of the reference gas stream to that of the measuring gas stream to prevent fluctuations in the flux signal after rapid changes in the ambient air CO 2 -concentration.
Download figure to PowerPoint
Each chamber was equipped with a Pt 100 temperature sensor. Every 10 s temperature, flow rate, CO2 difference and absolute concentration were recorded for each chamber and registered as a mean value over a 3 min sampling period.
Most of the direct measurements of rhizomicrobial respiration reported in the literature were performed with potted plants in the laboratory (Lambers et al., 1991). Field measurements are difficult because roots can not be extracted from the soil without any disturbance. To minimize root disturbance, Gansert (1994, 1995) developed the subtraction method. He inserted roots together with the surrounding soil into a measuring chamber. After the measurement the root was removed and the respiration of the remaining soil was measured. The rhizomicrobial respiration was calculated as the difference between the total respiration and the respiration of the remaining soil. The weak point of Gansert's approach was that the respiration rates of the root and soil samples were measured directly after they were taken from the soil and put into the chamber which means a disturbance to the rhizomicrobial system.
The improvement of the method in this study was that tips of fresh fine roots (referring to the German terminology ‘Langwurzel’ or long-root, Büsgen, 1901) were inserted into small plastic containers (18 × 13 × 6 cm) filled with sand in April 1995 where they developed a fine root system during the next month. From July to October each container received a weekly addition of 100 ml nutrient solution (5.0 mMol Ca2+ l−1 as CaCl2, 2.5 mMol Mg2+ l−1 as MgSO4, 12.5 mMol K+ l−1 as K2SO4, and KH2PO4, 2.5 mMol P l−1 as KH2PO4, 2.5 mMol N l−1 as NH4NO3). When grown in sand containers, fine roots had a different pattern of mycorrhizal colonization compared to the Of-horizon; only few poorly structured morphotypes were observed but these were very abundant (K. Pritsch pers. comm.). Previous experiments with a comparable design showed that fine roots of alder grown in sand have lower values of specific root length (0.8–1.5 m g−1 compared to a 3-yr mean of 8.0 m g−1 in the Of-horizon) and were more branched. The ratio of the length of fine lateral roots to that of the main roots on which they were borne (see definitions below) was more than doubled in sand containers compared to long-term means of the Of-horizon (Jacobsen, 1993; Middelhoff, 2000).
For the respiration measurements the containers were inserted into measuring chambers placed on the soil surface. To avoid an increase of the apparent respiration rate due to a CO2 concentration lower than that of the soil air (Palta & Nobel, 1989; Qi et al., 1994; Burton et al., 1997), the CO2-content of the measuring air stream, which was pressed continuously through the chambers, was held constant at 1500 ppm by means of a gas mixing unit (Walz, Effeltrich, Germany). To avoid drying effects in the chamber the air was passed through a bottle with distilled water. During the measurement the temperature inside the chambers was monitored with a Pt 100 at the surface of the root containers. The air in each channel was measured subsequently at 5 min intervals in the differential mode against air from the gas mixing unit. CO2 difference, absolute CO2 content of the air, temperature in and flow through the chambers were measured every 10 s and recorded as a mean value over the sampling period. After each measurement the root system was removed for further analyzes (see below) and the substrate was measured separately. Fig. 4 shows typical data sets for two containers measured from 10 to 14 October 1995. During the first 48 h of the experiment the containers included roots and sand. At noon on 12 October the roots were removed. Thereafter, the CO2 efflux of the containers was very small. Since the scatter of this basal respiration rate was very high it was not possible to fit temperature and moisture response curves for each single container. Thus the data of all measurements were pooled to fit a common correction term of the basal respiration, which was subtracted from the total container respiration to calculate the respiration of the rhizomicrobial system. At noon on 11 October a disturbance in the gas mixing unit caused a high scatter of the respiration values. These measurements were removed from the data-set. Finally, 811 data sets measured at 14 root containers were used for the data evaluation. Parallel measurements of roots recently extracted from the soil resulted in an up to six-fold higher respiration rate. This shows that the original approach of Gansert (1995) can not be applied to this organic soil.
Figure 4. Diurnal courses of the CO 2 -efflux from two of the root containers in October 1995. During the first part the containers contained an intact rhizomicrobial system, which was removed after 2 d of measurement. After that the sandy substrate was measured alone for 1.5 d. The arrows mark a period with a disturbance inside of the gas mixing unit, which caused fluctuations in the signal. Closed circles, root container 1; open circles, root container 2.
Download figure to PowerPoint
Definitions We distinguish two morphological and functional types of fine roots (< 2 mm). The woody, more elongated root type that evolves after secondary growth out of long roots is called ‘main-roots’ (also ‘mother roots’, Wilcox, 1964 ). They carry short-lived branches that are almost entirely mycorrhizal ( Pritsch, 1996 ) and in alder have only two orders of branching. The biomass relation [g g −1 ] between these lateral-roots and the main-roots is introduced as a measure of the branching intensity (BI).
When the soil respiration measurements in the alder forest were finished the soil within the frames was excavated to a depth of 20 cm and the mass of all roots with a diameter less than 5 mm and of the nodules was determined. After the direct measurements of the rhizomicrobial respiration the root system was removed from the container and main-roots and lateral-roots that had developed during the incubation were separated. Length and diameter of the main-roots was measured with a vernier gauge, the total length of the lateral-roots according to the ‘line intersection method’ (Tennant, 1975), which has an uncertainty of 5% (Tennant, 1975, own tests). D. wt of the main-roots and the lateral-roots were measured separately after drying at 65°C. Respiration rates were related to d. wt, total length, surface and volume of the roots gaining the best correlation against total root length (r2 = 0.45 for 110 measurements between 14° and 16°C).