Agroforest Syst DOI 10.1007/s10457-006-9006-9 Potential of nine multipurpose tree species to reduce saline groundwater tables in the lower Amu Darya River region of Uzbekistan Asia Khamzina Æ John P. A. Lamers Æ Christopher Martius Æ Martin Worbes Æ Paul L. G. Vlek Received: 25 May 2005 / Accepted: 24 April 2006 Springer Science+Business Media B.V. 2006 Abstract This paper evaluates the potential of nine multipurpose tree species for afforestation of degraded land in the Khorezm region, Central Asia, particularly their suitability for biodrainage i.e., lowering the elevated groundwater table through the transpirative capacity of plantations. For this purpose water use (WU), water use efficiency (WUE) and tree physiological factors influencing transpiration were assessed during two consecutive years. Mean daily leaf transpiration differed significantly among the species and ranged during the seasons from 4.5–5.2 mmol m–2 s–1 for Prunus armeniaca L. to 4.5–10 mmol–2 s–1 for Elaeagnus angustifolia L. WU differences were triggered by species physiological features such as capability of water uptake by roots. Transpiration rates and the length of fine roots correlated highly (r = 0.7). Correlations of leaf transpiration rates with leaf area were weaker (r = 0.6). No correlations were found between salt content in plants and water uptake under conditions of slight-tomoderate rootzone soil salinity. Values of WUE per root and shoot DM were similar averaging, respectively, 0.2 and 0.3 g DM g–1 water for twoyear-old trees, and decreased with age. In addition to WU characteristics, also salinity tolerance, growth rate and the ability to produce fodder and fuelwood must be considered during species selection. Regarding these features, the N-fixing E. angustifolia ranked the highest, combining high WU, fast growth and production of nutritious feed. Examined Populus spp. and Ulmus pumila L. ranked lower but still represented potential candidates for biodrainage purposes. Typical fruit species in the region such as P. armeniaca and Morus alba, showed low biodrainage potential. A. Khamzina (&) Æ J. P. A. Lamers Æ C. Martius Æ P. L. G. Vlek Department of Ecology and Resource Management, Center for Development Research, Walter Flex Str., 3, Bonn 53113, Germany e-mail: [email protected] Introduction M. Worbes Institute of Agronomy in the Tropics, University of Goettingen, Grisebachstrasse 6, Goettingen 37077, Germany Keywords Biodrainage Æ Root growth Æ Transpiration per unit leaf area Æ Water use efficiency High groundwater table (GWT) and salinity levels are the conditions prevailing in many irrigated landscapes of the Aral Sea Basin (Ibrakhimov et al. 2004; Martius et al. 2004). Khorezm region, an administrative district in the Republic of Uzbekistan, located at the southern edge of the Aral Sea Basin, is an oasis of irrigated farmland. Khorezm’s 123 Agroforest Syst landscape is characterized by insignificant slopes, low hydraulic conductivity in the upper soil horizons, and extremely slow lateral groundwater flows (Katz 1976; Mukhammadiev 1982). These adverse natural drainage conditions, aggravated by excessive irrigation and poorly maintained drainage systems, have resulted in the elevated GWT (Ibrakhimov et al. 2004). The evaporation of the shallow saline GWTs has caused secondary soil salinization throughout the entire irrigated area of Khorezm (Martius et al. 2004). Biological drainage uses the transpirative capacity of vegetation and especially trees, to cope with elevated GWTs in the landscape by enhancing their discharge or/and reducing their recharge. Since conventional drainage systems require expensive capital investments for installation, operation and maintenance, the implementation of biodrainage may represent a cost effective addition or even alternative (Heuperman et al. 2002). Extensive Australian research (e.g., Heuperman et al. 2002) concluded that the biodrainage plantations in the GWT discharge areas, which are e.g. prevailing in Khorezm, would not be effective on a regional scale. However, even the localized impact of biodrainage plantations is important for improvement of the degraded patches of land which are abandoned from agricultural activities and, being dispersed throughout Khorezm, currently constitute about 20% of the irrigated area in the region (Martius et al. 2004). Designing efficient biodrainage plantations to enhance the discharge of shallow and saline GWTs involves a careful selection of suitable tree species with preference to those having a high transpiration capability (Heuperman et al. 2002). The in-depth overview on biodrainage by Heuperman et al. (2002) emphasized the need to include, in addition to water use (WU) other criteria for tree selection such as salinity tolerance and, in particular, rooting characteristics. Among the latter, deep rooting, significant horizontal root system extension and increased root density indicated suitability for biodrainage purposes. Moreover, Heuperman et al. (2002) advocated the inclusion of site-specific characteristics thus stressing the need for field studies under natural conditions. The assumption that WU per unit leaf area (LA) differs among tree species has been debated 123 for a long time. Given the lack of evidence that stomata behavior is different among tree species even in conditions of water scarcity, Landsberg (1999a) concluded that water use per unit LA, is not an appropriate indicator for the choice of species. In contrast, Deans and Munro (2004) examined WU of dry land trees such as Acacia seyal Del and Acacia aneura F. Muell ex Benth and revealed significant differences in leaf transpiration rates among species. Existing knowledge on tree WU in Uzbekistan is insufficient since transpiration rates of various locally grown tree species were estimated (e.g., Botman 1988; Mikhailova 1986) solely by using destructive gravimetric methods measuring weight differences over periods of 3 min (Ivanov et al. 1950). To test the hypothesis whether transpiration rates per unit LA are the same irrespective of tree species, the present study examined the characteristics of nine native and introduced tree species under the agro-ecological conditions of the Aral Sea region, using Khorezm as a model region. An effort was made to identify those species which combined a high WU with salinity tolerance and an ability to rapidly produce high quality fodder and fuelwood (Khamzina et al. 2006), as this may increase farmers’ willingness for adopting the low-cost biodrainage technology. Materials and methods Description of the study sites The research was conducted at the Khiva Research Station of the Uzbek Forestry Research Institute located at 4141¢ N latitude, 3940¢ E longitude and at an altitude of 113 m. The study region is characterized by an arid, strongly continental climate with high irradiance and sparse precipitation. Average annual precipitation between 1990 and 2002 amounted to 101 mm and was exceeded by precipitation for the study years 2002 and 2003 by 75% and 70%, respectively. Figure 1 depicts the experienced atmospheric parameters that govern the loss of water via leaf transpiration. Experimental plantations were established at two sites which measured 0.14 ha each and were Agroforest Syst b) middle (16 MaP) 50 40 1200 30 800 20 400 10 1200 30 800 20 400 0 d) end (7 MaP) 2000 60 50 1600 50 40 1200 30 800 20 400 10 0 0 9-10 11-12 13-14 15-16 17-18 19-20 Time, h 2000 1600 40 1200 30 800 20 PFD, µmol m-2 s-1 60 PFD, µmol m-2 s-1 T, °C & RH, % 40 0 c) end (18 MaP) 7-8 1600 50 10 0 0 2000 60 1600 T, °C & RH, % T, °C & RH, % 60 70 T, °C & RH, % AT RH PFD 2000 PFD, µmol m-2 s-1 70 PFD, µmol m-2 s-1 a) beginning (14 MaP) 400 10 0 0 7-8 9-10 11-12 13-14 15-16 17-18 Time, h Fig. 1 Air temperature (T), relative humidity (RH) and photon flux density (PFD) in the beginning (14 months after planting (MaP)), middle (16 MaP) and the end (7 and 18 MaP) of the growing seasons 2003–2004 located 500 m apart from each other. Soil profiles represented two major soil textures in Khorezm: (i) a light sandy soil underlain by a loam layer from 75 cm downwards (hereafter referred to as the sandy site) and (ii) a more finely textured silt loam underlain by a loam layer from 85 cm downwards (loamy site). The available water content estimated as the difference between soil moisture at the field capacity and at the wilting point, determined by the pressure membrane method, averaged to 89 and 49 mm m–1 within the examined profiles at the loamy site and sandy site, respectively. Both sites were leveled and leached from salts in the beginning of the study period, which provided homogeneity of initial growing conditions. More details on soil properties and methods for soil analysis are reported in Khamzina et al. (2006). Due to adequate availability of irrigation water in the region during the study years, the presence of intensively irrigated rice fields in the vicinity, and a high GWT, the trees at the loamy site only needed occasional irrigation. At the sandy site, where the soil moisture remained close to field capacity, only one irrigation event was necessary at the onset of each growing season. For the rest of the time the trees relied on groundwater and precipitation. The GWT at the sandy site was in the range of 0.7–1.3 m during the growing period, whereas at the loamy site it ranged from 1.0 to 1.3 m. Thus, the average GWT depth monitored at the sites was equal or less than the long-term mean of 1.22 m for the region ((Ibrakhimov et al. 2004); only those months observed in the sites were used for comparison). The GWT dropped during the winter and rose again with the start of spring leaching in the surrounding agricultural fields. During both observation years mean electrical conductivity (EC) of GWT was highest at the 123 Agroforest Syst non-irrigated sandy site (4.3 dS m–1 vs. 3.3 dS m–1 at the loamy site). Both exceeded the long-term mean salinity of 1.8 g l–1 (Ibrakhimov et al. 2004). Experimental design The experiment was laid out in a randomized incomplete block design with six replications. Plots consisted of two rows with 25–30 trees, spaced 1 m · 3.5 m. This spacing was introduced to minimize potential root competition between neighboring trees and to avoid mutual shading during the experiment. The pre-selected ten deciduous species represented a variety of life spans and tolerances to drought and salt stress: apricot tree (Prunus armeniaca L.), black poplar (Populus nigra var. pyramidalis (Rozan) Spach), black willow (Salix nigra Marshall), Eastern catalpa (Catalpa bignonioides Walter), Euphrates poplar (Populus euphratica Oliv.), Russian olive (Elaeagnus angustifolia L.), salt cedar (Tamarix androssowii Litv.), Siberian elm (Ulmus pumila L.), swamp ash (Fraxinus pennsylvanica Marshall), and white mulberry (Morus alba L.). Seeds of all species were collected from plantations on the station, germinated and grown for one year in a tree nursery located 7.2 km from the experimental fields. After 12 months, in total 50 saplings of each species were transplanted into the experimental plots. At each site, the tree species were completely randomized within the experimental plots. Transpiration rates and leaf area Leaf transpiration rates of eight trees per species were measured with a steady state porometer (Li-cor 1600). Concurrently, the porometer sensors measured irradiance (photon flux density), air temperature and relative humidity. The first porometer surveys occurred 7 months after transplanting the saplings from the nursery into the experimental plots (MaP). This corresponded with the end of the first growing season, when the leaf area became sufficiently large to fit the narrow leaf aperture cap. Subsequent measurements were conducted three times during the second growing season each time for five consecutive 123 days: shortly after the leaves flushed, in the mid-season and at the end of the season, which corresponded to 14, 16, and 18 MaP. At each tree, the abaxial side of three-to-four sunlit leaves was measured every 2 h during 13 hours of sunshine duration. The fine and scaly leaves of T. androssowii were unsuitable for measurements with the available aperture cap. Also, P. euphratica at the loamy site developed very narrow leaves inappropriate for the aperture cap. Finally, due to time and equipment limitations, during the second growing season at the sandy site only three species could be measured at 14, 16, and 18 MaP. These had shown high growth and utility potential at 7 MaP. At 7 and 19 MaP (the end of the two growing seasons in 2002–2003) the observed trees were completely defoliated and fresh leaf mass determined. Leaf sub-samples were transported in a cool-box to the laboratory to determine total oneside LA in two replicates with a leaf area meter (Li-cor 1200). Daily transpiration of the whole canopy was estimated using total LA of the measured trees and assuming no self-shading within canopies. Water use efficiency (WUE) of trees was calculated as the estimated canopy transpiration with relation to below- and aboveground dry matter (DM) production according to Larcher (1995, pp. 121 ff). Biomass At 7 and 19 MaP, 3–4 trees per species and plot were cut at ground level and separated into stem, branch and foliage. P. euphratica was not harvested at 7 MaP at the loamy site due to low survival rate. Complete tree root systems were manually excavated to quantify belowground biomass and rooting extent. Maximal rooting depth and maximal radial extension of the root system were determined with a measuring tape. In the laboratory, roots were washed free of soil and separated into coarse (Ø >3 mm) and fine roots (Ø < 3 mm). Coarse root length was measured with a measuring tape. The length of live fine roots was determined using the modified Newman lineintersect method (Tennant 1975). All above- and below-ground fractions were dried to constant weight at 103C to determine DM. The relative Agroforest Syst relationship of the above- and below-ground DM was expressed as root/shoot ratio (RSR). Chemical analysis The concentration of salts in plant tissues, particularly leaves, provides information about salt accumulation ability and salt tolerance of plants (Heuperman et al. 2002). Therefore, dried samples from leaves, twigs, stems, coarse, and fine roots were collected at 19 MaP when the accumulated salt content in plant tissues was expected to be the highest. The samples were selected from trees with a leaf mass judged representative for each species and soil type, ground and analyzed for contents of chloride (Cl–) and sulphate (SO2– 4 ), the dominating soil salt anions in the region. Concentration of salts in soil solution and tissues was determined on a 1:5 water extract. Chloride content was measured titrametrically. Sulphate concentration was determined by precipitation with 10% BaCl2. species and soil type and their interactions were compared at p < 0.05 level of significance. Individual treatment means were compared with the Tukey Post Hoc test where the ANOVA test had indicated significant effects. All statistics were carried out with SPSS 11.0 software. Results Aboveground biomass The GWT was monitored using four observation wells (Ø = 4 cm) installed down to a depth of 1.8– 2.2 m. The wells were polyethylene pipes closed at the bottom, perforated and protected from clogging with a fine synthetic filter. The EC of the groundwater was measured every 10 days as well as immediately before and after each irrigation event. At two locations at each site, in the beginning and end of each growing season soils were sampled in two replicates every 0.2 m horizons down to the GWT for the analysis of salt contents. The total DM production at 7 MaP was in general higher for all tree species at the sandy site, but at 19 MaP some trees at the loamy site had gained higher total DM. The utilizable (leaves and wood) aboveground DM at 7 MaP was also higher at the sandy site but at 19 MaP for a number of species this did not differ significantly between the two sites. The RSRs were consistently higher at 7 MaP than at 19 MaP especially for the sandy soil, but this soil-specific difference disappeared for most species at 19 MaP (Fig. 2 and Table 1). With regards to total DM production E. angustifolia, P. nigra var. pyramidalis and U. pumila, always ranked in the upper quartiles irrespective of soil type and harvest date. P. euphratica, though a good performer at the sandy site, was hardly productive in the loamy soil. S. nigra and M. alba, however, had a higher total DM production at 19 MaP at the loamy site and ranked in the upper quartiles. Whereas growth of a species such as U. pumila was shifted towards the belowground DM, other species such as E. angustifolia, invested relatively more in the aboveground DM, as suggested by its lower RSR (Fig. 2 and Table 2). Statistical analyses Root growth parameters All data were checked for normality and normalized using the logarithmic transformation when necessary. Analysis of variance (ANOVA) was performed using the General Linear Model procedure. With bivariate Pearson correlations the relations were tested between mean daily transpiration rate and environmental and physiological parameters measured at 7 and 19 MaP. The effect of the treatment variables such as tree The ANOVA revealed a highly significant influence of species, soil type and harvest date on virtually all root growth parameters (Tables 1 and 2). At 7 MaP the DM and, for the most part, total lengths of the fine roots were greatest for species at the sandy site (Fig. 3 and Table 1). The light soil texture and higher GWT apparently enhanced fine root growth. At 19 MaP these trends were maintained, although some species such as Groundwater and soil sampling 123 Agroforest Syst a) 7 MaP Coarse roots Stem Leaves 0.6 Table 1 Standard errors of the differences between means according to harvest dates (MaP) and soil types Fine roots Twigs Root/shoot ratio 2.0 Factor/parameter 1.2 0.3 0.8 0.2 sandy FP MA CB PA SN UP EA PN FP MA CB PA 0.0 PE 0.0 SN 0.4 EA 0.1 Root/shoot ratio 1.6 0.4 PN UP Biom ass, kg tree -1 0.5 loamy b) 19 MaP 5 1.2 2 0.8 0.4 0 0 sandy loamy Fig. 2 Dry matter production of nine tree species and root/shoot ratio according to the soil type at 7 months after planting (2a: 7 MaP) and 19 MaP (2b). Connecting lines of RSR are to aid in viewing trends only. Species are ranked in descending order according to the total biomass in sandy soil. CB = C. bignonioides, EA = E. angustifolia, FP = F. pennsylvanica, MA = M. alba, PA = P. armeniaca, PE = P. euphratica, PN = P. nigra var. pyramidalis, SN = S. nigra, UP = U. pumila P. armeniaca and C. bignonioides did not perform significantly worse at the loamy site. An exception occurred with fine root DM of M. alba, which was almost by 50% greater in the loamy than in the sandy soil at 19 MaP. Growth response of structural roots to the soil texture was not uniform. With age, M. alba, S. nigra and P. armeniaca developed considerably greater coarse root DM in loamy soil (Fig. 2). However, the length of coarse roots of these species did not differ significantly with soil type (Fig. 3 and Table 1). The other tree species developed higher structural root mass and length in sandy soil. Shallow GWT at both sites prevented root development below the capillary fringe. Even a 123 92.8*** 19.1*** 2.8*** 18.1*** 36.8*** 21.2*** 0.1** 127.8*** 176.5*** 4.8ns 10.2*** 75.3* 42.1*** 4.1*** – – Soil 97.1ns 20.1ns 2.6*** 18.9ns 36.4ns 21.6ns 0.1* 133.1ns 202.6** 4.3ns 9.3*** 84.8ns 48.3ns 5.1ns – – Soil 7 MaP 19 MaP 26.3* 9.2* 1.8*** 4.8ns 8.0ns 4.9ns 0.1* 31.3ns 33.7ns 5.8ns 6.8*** 170.7ns 94.4ns 0.5ns – 294.9ns 58.1ns 7.2** 59.0ns 119. 3ns 68.3ns 0.1ns 376.1* 491.7*** 2.4*** 24.7*** 7.1ns 11.8ns 7.3** 0.2* – 0.8ns *, **, *** respectively indicate significance levels < 0.05, < 0.01 and < 0.01; ns = not significant EA PE PN UP SN FP MA CB PA 1 Root/shoot ratio 1.6 3 EA PE PN UP SN FP MA CB PA Biom ass, kg tree -1 2 4 Total DM Coarse root DM Fine root DM Stem DM Twig DM Leaf DM Root/shoot ratio Coarse root length Fine root length Maximal depth Maximal radius WUE per root DM WUE per shoot DM Tree transpiration SO–2 4 content in all fractions Cl– content in all fractions MaP species such as U. pumila, which is genetically predisposed to develop deeply penetrating taproots (Forestry Compendium 2000), had no prominent primary root, but instead a rather dense and superficial lateral root system. Trees at the loamy site rooted in general more deeply at 19 MaP, due to the lower GWT (Fig. 3 and Table 1). The radial extension of coarse roots in sandy soil varied according to species and was considerable, particularly for both Populus spp. At the loamy site, the horizontal spread of the coarse roots was less significant but varied according to species at 7 MaP, with P. nigra var.pyramidalis exploiting the largest radius (Fig. 3 and Table 2). Among the examined root parameters, fine root length was highly and positively correlated with transpiration rates of the trees (Table 3). Although maximal rooting depth was not consistently different among the species, increased transpiration rates were correlated with the vertical expansion of the root systems. Tree species with considerable horizontal extension exhibited rather low root length densities (RLD), and therefore correlations between transpiration rate and RLD were poor and in most cases even Agroforest Syst Table 2 Differences between species means as presented in Figs. 2 and 3 Parameter p Sandy soil Loamy soil CB EA FP MA PA PE PN SN UP CB EA FP MA PA PE PN SN UP 7 MaP Total DM Coarse root DM Fine root DM Wood DM Leaf DM Root/shoot ratio Coarse root length Fine root length Maximal depth Maximal radius < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.150 0.032 < 0.001 < 0.001 < 0.001 abc abc ab a ab a a abc a ab d c bc d c a a bc a ab abc abc bc ab ab a a c a ab ab ab ab a a a a ab a ab a a a a a a a a a a cd bc ab bcd bc a a abc a b bcd abc ab cd ab a a abc a ab abc abc ab abc ab a a abc a ab abc abc c abc a a a bc a ab ab a abc a ab abc ab ab b ab bc ab bc bc bc a b ab b ab ab a bc ab ab c b ab ab ab ab a bc a ab abc ab ab b ab a a a a a ab a a a a – – – – – – – – – – c b bc c c ab b b b b ab a ab ab ab abc ab ab ab ab ab ab c ab ab bc b b b ab 19 MaP Total DM Coarse root DM Fine root DM Wood DM Leaf DM Root/shoot ratio Coarse root length Fine root length Maximal depth Maximal radius < 0.001 0.001 0.002 < 0.001 < 0.001 0.025 < 0.001 0.076 0.085 0.014 a a a a a bcd a ab a a b c ab b c a b abc a abc a abc ab a ab d ab bc a a a ab ab a ab abc ab abc a ab a a a a a bc a a a a ab bc ab ab bc ab b abc a bc a abc b a ab abc ab c a c a ab ab a ab ab ab abc a abc a abc ab a ab cd ab bc a abc a a a a a a a ab a a b ab abc b b a b abc a a a a ab a a a a abc a a ab ab bc a a a ab b a a a a a a a a a abc a a a a a a a a a a a a ab ab abc a ab a ab abc a a ab b abc a ab a ab abc a a ab ab c a a a ab bc a a The same superscripts within the columns indicate that the means are not significantly different at p < 0.05. The species names are spelled out in Figs. 2 and 3 negative. Correlation coefficients between transpiration and total tree LA were mostly high but did not indicate a consistent correlation (Table 3). Despite differences in species performance with respect to soil type, harvest date and parameter measured, species such as E. angustifolia, both Populus spp. and U. pumila were found in the upper quartiles for virtually all rankings. Among these, P. euphratica was the leading species at the sandy site but showed slow root elongation and low DM production at the loamy site. On the other side of the spectrum after all rankings was P. armeniaca, which always had the lowest belowground DM irrespective of the tested parameters and soil type (Table 2). Transpiration The diurnal transpiration courses of most species showed one daily peak at different points during the day depending on the period in the season (Fig. 4). At 14 MaP the highest values recorded at the loamy site happened between 11:00 and 12:00 h. At 16 and 18 MaP water loss values at the loamy site peaked between 13:00 and 14:00 h and for U. pumila, M. alba and S. nigra additionally between 17:00 and 19:00. These three species decreased their transpiration during the hottest time of the day. At the sandy site, during the early, mid and late season, all species peaked at 11:00 and 12:00 h, and at the end of the season a second peak was observed between 15:00 and 16:00 h. The transpiration rates per unit LA were much higher at 18 MaP than at the end of the previous season (Table 4). Mean values of leaf transpiration followed the sequence 16 MaP (mid growing season) 18 MaP (late season) > 14 MaP (early season) > 7 MaP (end of the first season), which was consistent with the mean air temperature and relative humidity during these periods. On the other hand, this sequence did not correspond with the irradiance given by in average higher values of photon flux density (PFD) observed at the end of both growing seasons. 123 Agroforest Syst Fine root length Coarse root length Ground water table Maximal rooting depth Maximal rooting radius*10 16 14 0.0 0.2 12 0.6 8 0.8 6 Depth, m 0.4 10 1.0 4 1.2 0 1.4 EA CB UP PN FP SN MA PA 2 EA CB UP PN PE FP SN MA PA Root le ngth and ra dial e xte ns ion, m a) 7 MaP sandy loamy 120 0.0 100 0.2 0.6 60 0.8 40 Depth, m 0.4 80 1.0 1.2 0 1.4 EA PN PE UP FP SN MA CB PA 20 EA PN PE UP FP SN MA CB PA Root le ngth and ra dial e xte ns ion, m b) 19 MaP sandy loamy Fig. 3 Length of coarse and fine roots, maximal radius and depth of root systems of nine tree species, and GWT depth according to soil types and harvest dates. Values of max radius are multiplied by 10 for improved visualization. Connecting lines of max radius and depth are to aid in viewing trends only. Species are ranked in descending order according to the total length in sandy soil. CB = C. bignonioides, EA = E. angustifolia, FP = F. pennsylvanica, MA = M. alba, PA = P. armeniaca, PE = P. euphratica, PN = P. nigra var.pyramidalis, SN = S. nigra, UP = U. pumila ANOVA revealed a significant effect of species on mean transpiration rate per unit LA at the loamy site (Table 4). At the sandy site, where only three the best performing species could be measured during the second growing season, species-related differences were found only at the onset of the season with E. angustifolia transpiring significantly more water. Transpiration of trees grown in the loamy soil was significantly higher than that of trees at the sandy site but at 7 and 16 MaP only. Overall, the highest water consumption per unit LA was shown by E. angustifolia, U. pumila and P. euphratica at 18 MaP whereas P. armeniaca was the least transpiring species during this period of the season. Water use efficiency WUE in physiological studies is usually measured as the ratio of photosynthetic to transpiration rate, but for ecological, agricultural, and forestry purposes the relation of DM production to water consumption is considered more informative (Larcher 1995, pp. 121 ff; Landsberg 1999b). The WUE was calculated separately for above(WUEa) and belowground (WUEb) DM production. Overall WUE varied considerably between species at 19 MaP. In particular, the WUE values of P. armeniaca at the sandy site and of F. pennsylvanica at the loamy site at 7 MaP by far exceeded those of the other species, whereas this trend was not maintained at 19 MaP. The values of WUE per unit root DM production Table 3 Pearson correlation coefficients for mean daily transpiration rate, leaf area, fine root DM, coarse root DM, fine root length, coarse root length, maximal radius and maximal depth of root systems, root/shoot ratio, and root length density Species (N of cases) Leaf area Coarse Fine Root/shoot Coarse Fine Maximal Maximal Root length root DM root DM ratio root length root length depth radius density C. bignonioides (5) 0.87 0.52 E. angustifolia (11) 0.80* 0.69* F. pennsylvanica (8) 0.54 0.52 M. alba (7) 0.61 0.65 P. euphratica (6) 0.71 0.58 P. nigra var. pyramidalis (5) 0.88 0.91* P. armeniaca (6) 0.91* 0.82* S. nigra (4) – 0.91 U. pumila (12) 0.78* 0.73** Overall 0.60** 0.62** 0.02 0.53 0.05 0.44 0.81 0.50 0.59 0.68 0.62* 0.61** –0.71 –0.82** –0.48 –0.89** –0.90* –0.91* –0.45 0.55 –0.76** –0.54** 0.77 0.59 0.46 0.53 0.58 1.00** 0.43 0.91 0.34 0.51** 0.90* 0.60 0.85** 0.84* 0.81 0.96* 0.56 1.00* 0.66* 0.70** 0.93* 0.59 0.92** 0.85* –0.42 0.93* 0.85* 0.98* –0.40 0.21 *, **, *** indicate correlations respectively significant at 0.05, 0.01 and < 0.01 level (2 tailed) Numbers in bold indicate significant correlations 123 0.03 0.58 0.61 0.32 0.72 0.52 0.86* 0.51 0.64* 0.51** 0.18 –0.57 –0.95** 0.62 –0.49 –0.66 –0.44 0.71 –0.61* –0.21 -2 Transpiration, mmol m s -1 Agroforest Syst 16 14 12 14 MaP (early season) 16 MaP (mid season) 18 MaP (late season) a) Sandy soil 10 8 6 4 E.angustifolia 2 M.alba P.euphratica 0 -2 Transpiration, mmol m s-1 16 P.nigra P.armeniaca b) Loamy soil 14 S.nigra U.pumila 12 10 8 6 4 2 0 7-9 9-10 11-12 13-14 15-16 17-18 19-20 Time, hrs 7-9 9-10 11-12 13-14 15-16 17-18 19-20 7-9 9-10 11-12 13-14 15-16 17-18 19-20 Time, hrs Time, hrs Fig. 4 Diurnal variations of leaf transpiration rates at the beginning (14 MaP), middle (16 MaP) and end (18 MaP) of the growing season according to tree species and soil type Table 4 Mean daily leaf transpiration rates (mmol m–2 s–1) at the beginning (14 MaP), middle (16 MaP) and end (7 and 18 MaP) of the growing seasons according to tree species and soil type Species (N of cases) 7 MaP Sandy site C. bignonioides (42) E. angustifolia (53) F. pennsylvanica (48) M. alba (52) P. armeniaca (48) P. euphratica (48) P. nigra var. pyramidalis (44) S. nigra (44) U. pumila (52) 2.11a 2.11a 2.46a 2.22a 2.20a 2.22a 2.43a 2.30a 2.28a – 4.51b – – – 3.30a – – 3.30a – 5.12a – – – 4.82a – – 4.53a – 9.99a – – – 8.32a – – 9.22a Loamy site C. bignonioides (106) E. angustifolia (171) F. pennsylvanica (118) M. alba (108) P. armeniaca(69) P. nigra var. pyramidalis (99) S. nigra (115) U. pumila (150) 2.12a 2.68b 2.36ab 2.79b 2.40ab 2.50ab 2.40ab 2.62b – 4.37b – 2.64a – – – 3.63b 4.38ab 9.12c 7.13bc 6.51ab 5.15ab 2.99a 7.31bc 7.38bc 6.69ab 9.39b 7.86ab 7.46ab 4.48a 6.04ab 5.70ab 9.49b Analysis of variance, probability > F (=alpha) Soil 0.002 Species 0.239 Species*soil 0.079 0.699 < 0.001 0.323 < 0.001 < 0.001 0.266 0.773 < 0.001 0.436 Across harvest dates MaP Soil Species Soil*species 14 MaP 16 MaP 18 MaP < 0.001 < 0.001 < 0.001 < 0.001 Means with the same superscript within the column are not significantly different at p < 0.05 123 Agroforest Syst were higher than those per unit aboveground DM production in both species. At 19 MaP, P. armeniaca and S. nigra were the best performers regarding WUEa, and F. pennsylvanica was the most efficient species regarding WUEb (Fig. 5). WUE significantly decreased with tree age (Table 1): the most for F. pennsylvanica, the least for E. angustifolia and P. nigra var. pyramidalis. A general trend at 7 MaP was that WUE was inversely related to daily canopy transpiration. At 19 MaP, this trend was not observed (Fig. 5). WUE per root DM WUE per shoot DM Transpiration WUE, m g DM g-1 water 1600 5 1400 4 1200 1000 3 800 2 600 400 1 200 Transpiration, l d -1 tree -1 a) 7 MaP 0 0 A E P P B N A PAM P F U C P E sandy PA MA FP UP CB PN EA loamy b) 19 MaP Salt accumulation in plant tissues Average soil salinity within the top 1 m layer, determined by a sum of potentially toxic ions (Kovda et al. 1985) was classified as slight-tomoderate. At both sites, the prevailing salinity type was defined as sulphate-chloride. The soils at the sandy site had significantly higher sulphate salt contents than those at the loamy site (Table 5). The EC of the GWT was also higher at the sandy site. Consequently, the content of SO2– 4 was higher in plant tissues of trees grown on the sandy soil when compared to those on the loamy soil (Fig. 6 and Table 1). In contrast, Cl– content did not differ according to the soil type. The mineral content varied depending on species but particularly the Cl– content at the sandy site did not differ significantly between species. Salt concentrations highly depended upon the plant fraction (p < 0.001). In line with previous observations that responses to soil salinity occur first in the shoot, despite the root being exposed to the soil salinity (Poljakoff-Mayber and Lerner 1999), the highest Cl– concentrations were found in leaves and fine root fractions rather than in the perennial parts, whereas SO2– 4 was mostly concentrated in tree leaves and stems. WUE, m g DM g-1 water 60 160 50 120 40 30 80 20 40 10 Transpiration, l d -1 tree -1 200 0 0 UP EA PE sandy UP EA PA SN PN FP MA CB loamy Fig. 5 Water use efficiency per root and shoot dry matter (WUE) and daily tree transpiration of nine tree species according to soil types and harvest dates. Species are ranked in descending order according to the total WUE in sandy soil. Species effect is significant for WUE per shoot and root DM (p < 0.01) at 19 MaP only and for transpiration both at 7 MaP (p < 0.001) and 19 MaP (p < 0.01). CB = C. bignonioides, EA = E. angustifolia, FP = F. pennsylvanica, MA = M. alba, PA = P. armeniaca, PE = P. euphratica, PN = P. nigra var. pyramidalis, SN = S. nigra, UP = U. pumila 123 Table 5 Concentration of Na+, Cl– and SO–2 ions 4 (cmol kg–1) in 1 m soil layer in the beginning and end of the growing seasons Sampling date Na+ Cl– SO–2 4 Sandy site 1 MaP 7 MaP 13 MaP 19 MaP 0.6a 1.1a 1.1a 1.0a 0.4a 0.8a 0.7a 0.5a 1.6a 1.9a 2.0a 2.4a Loamy site 1 MaP 7 MaP 13 MaP 19 MaP 0.7a 1.3b 1.1ab 0.9ab 0.4a 0.7b 0.7b 0.5ab 0.8a 1.7b 1.3ab 0.9a 0.01 0.816 0.996 0.031 0.06 0.749 ANOVA, probability > F (=alpha) MaP 0.687 Soil 0.632 MaP*Soil 0.417 Means with the same superscript within the column are not significantly different at p < 0.05 Agroforest Syst The sulphate content in tissues of all tree species was significantly lower than that of the halophytic shrub T. androssowii which is known for its ability to accumulate salts (Forestry Compendium 2000). The chloride content in C. bignonioides, U. pumila and especially F. pennsylvanica at the sandy site was relatively high and was comparable with that in T. androssowii (Fig. 6). Transpiration rates did not appear to correlate with concentration of the examined salts in the plant tissues measured at the end of the growing season 2003. Discussion Loamy_leaves Sandy_leaves Loamy_all Sandy_all 7.5 5 4 6 3 4.5 2 3 1.5 1 0 0 3 1 2.5 0.8 2 0.6 1.5 0.4 1 0.2 0.5 0 SO4- 2 in all fra ctions, g 10 0 g- 1 D M 9 Cl- in all fra ctions , g 10 0 g- 1 DM Cl- in leaves, g 100 g -1 DM SO4-2 in leaves, g 100g -1 DM The choice of tree species is crucial when considering trees as a means to enhance discharge of 0 PA CB EA FP MA PE PN SN UP TA Species Fig. 6 Mean sulphate and chloride content in leaves and that averaged over all plant fractions (leaves, twigs, stems, fine and coarse roots) for nine tree species at 19 MaP according to the soil type. Species are ranked in ascending order according to sulphate content in leaves at the sandy site. Species and soil effects are significant for sulphate content (p < 0.001 and p < 0.01, respectively). CB = C. bignonioides, EA = E. angustifolia, FP = F. pennsylvanica, MA = M. alba, PA = P. armeniaca, PE = P. euphratica, PN = P. nigra var. pyramidalis, SN = S. nigra, UP = U. pumila, TA = T. androssowii the shallow and saline GWTs within degraded patches of irrigated land in the Aral Sea Basin. The ideal multipurpose species should combine a number of features such as high survival rates, quick growth, halophytic and xerophytic characteristics, and high utility value of firewood or foliage (Khamzina et al. 2006) since financial returns are key to farmers (Landsberg 1999a). For biodrainage purposes, additional characteristics such as WU and salinity tolerance should be included as selection criteria. We have examined such parameters of nine deciduous tree species and attempted to correlate WU of the trees with the internal factors, determined by physiological features of the tree species, and the environmental parameters. Influence of environmental factors on the transpiration As the soil moisture conditions were maintained at an adequate level during the experiment, we assume that water uptake by roots met the requirements of potential evapotranspiration imposed by atmospheric factors. The dependency of transpiration on various atmospheric variables makes it difficult to find a clear relationship on a single one of these (Angelocci et al. 2004). This was also experienced in this study, given the consistent—but not high—dependency of transpiration rates over all species on air temperature (r = 0.57), relative humidity (r = 0.47) and irradiance (r = 0.47). These relationships improved when differentiated according to species, since the tree-specific physiological features (such as ability to regulate stomata opening) together with environmental factors altered the transpiration rates. Diurnal transpiration curves showed unrestricted water loss of some species such as E. angustifolia, while M. alba, S. nigra and U. pumila were more sensitive to high temperature and irradiance and demonstrated a reduced transpiration during the hottest time of the day, which decreased the mean daily transpiration by these species. Presence of salts in soil and water can significantly reduce WU of plants (Poljakoff-Mayber and Lerner 1999), which, for the hydromorphic salt-affected soils in Khorezm makes salt 123 Agroforest Syst tolerance an important criterion for species selection. The level of soil salinity experienced at the experimental sites, however, did not affect the growth and transpiration of most of the examined species, which were able to expel salts from the ‘‘single phase water uptake’’ (Heuperman et al. 2002). This may explain the low correlation between leaf transpiration rates and concentrations of predominant salts accumulated in the tissues and generally the low salt concentration in tissues of the tree species in comparison with halophytic T. androssowii. However, the sandy site had higher values of groundwater EC than the loamy site which clearly affected P. armeniaca, C. bignonioides, F. pennsylvanica and M. alba—all species that previously were reported as being salt sensitive (Fimkin 1983). The transpiration of these species could not be measured at the sandy site in the second season but the observed visual signs of stress (leaf chlorosis and necrosis) correspond with lower transpiration rates. On the other hand, high transpiration rates of E. angustifolia and U. pumila at the sandy site fall in the range of their transpiration values at the loamy site, showing that these species were rather insensitive to the present degree of groundwater and soil salinity. Influence of tree physiological factors on transpiration The high correlations found between the transpiration rates per unit LA and root system characteristics and morphology are in line with other findings. Bi et al. (1992) studied the competitive advantages of young Eucalyptus trees and pointed out the importance of measuring the RLD. So did Nnyamah and Black (1977), who reported a good relationship between water uptake and rooting density. However, the directly measured root length in this study was more indicative than RLD. This may be due to the fact that RLD was an estimate over the whole rooting profile using root length, and radial and vertical extensions of root systems, thus not taking into account the root stratification within soil horizons. Theiveyanathan and Benyon (2000) associated the relatively high WU of Corymbia 123 maculata with the ability to develop high root densities in the area of the capillary fringe. The capacity of the root system of Eucalyptus spp. to adapt to local environment was witnessed by Knight (1999) who associated the ability of these species to maintain a relatively high transpiration rate over dry periods to their dimorphic root systems. The eucalypts combined deep roots for taking up the groundwater with superficial lateral roots for rainfall interception. The prevalence of high GWTs in Khorezm prevents trees to develop taproots, as these may decay in conditions of elevated GWTs. Local and introduced species (for the most part early last century) tended to expand their roots horizontally and thus developed a substantial length, particularly of fine roots, which are the primary assimilators of water. However, the tendency for horizontal root spread may result in lower root densities, which can explain weak and negative correlations of RLD and transpiration rates of the trees. The relationship of transpiration rates with LA was less important (r = 0.56) than that with root length (r = 0.70), at least during the vital establishment phase. This does not correspond with findings of Hatton et al. (1998); who, however, measured transpiration of evergreen tree species in closed stands. We suppose that under the agroclimatic conditions of the study region the root formation was more essential during the early growth of the young trees investigated, although at a later stage the importance of LA is expected to increase. Also, deciduous trees which produce a new canopy each season cannot adjust their leaf areas constantly and rapidly (Landsberg 1999a) while the fine root turnover is very high (e.g., Santantonio 1990). The variability in transpiration rates per unit LA among tree species even under ample water availability does not support findings of Hatton et al. (1998) who postulated that transpiration rates were independent of species. Yet, even when water stress, which alters stomata behavior of drought-sensitive species, is excluded, speciesdependant physiological features, such as tolerance to salinity and poor drainage, do influence transpiration rates. Naturally co-occurring species may have come to the same long-term solution to the water and salt stress and develop similar Agroforest Syst transpiration rates. However, all trees in the agricultural area of Khorezm are deliberately planted and therefore likely to retain their unique adjustment mechanism. These outcomes emphasize the strong influence of species factor on WU and therefore suggest that direct measurement of leaf transpiration of tree species planted in salt-affected agricultural land in Khorezm is preferable over the use of speciesindependent micrometeorological techniques; in particular, since the latter may not be effective for the estimation of evapotranspiration in heterogeneous stands (e.g., Schaeffer et al. 2000). The findings at the sandy field suggest that in years of sufficient water availability it is possible to establish tree plantations with little surface irrigation but with reliance on the GWT. However, given the strong dependency of GWTs on the agricultural irrigation activities in the direct surroundings (Ibrakhimov et al. 2004) and the normally meager rainfall during the growing season, relying on GWT alone for tree establishment may be insecure. Only after root system establishment, trees can be expected to survive solely on groundwater supply and have any biodrainage effect. Another problem related to the sustainability of biodrainage plantations is that GWT salinity is likely to rise if irrigation is discontinued, which adversely affects the most salinity-sensitive species—a problem that yet needs consideration if biodrainage is to be successful. Water use efficiency Many WUE studies involving trees considered aboveground DM (wood plus leaves) only. This was often due to constraints associated with quantifying the root fraction although some studies successfully modeled root DM (e.g., McMurtrie et al. 1990). When ranking our species according to their WUE, the sequence substantially changed depending on whether the considered DM component was the above-, the below-ground biomass fraction or both. This finding bears out the importance of incorporating the below-ground fractions in tree WUE estimates in particular when WUE is an intended criterion for species selection. However, when water is not a limiting factor the species with a maximal WU such as E. angustifolia must be prioritized for purposes of biodrainage, even if their WUE is low. This, therefore, restricts the use of WUE as a selection factor for trees, which are considered for reducing high GWTs. Advocates of biological drainage plantations often emphasize the low-cost aspects for the removal of excess water (Smedema 1997). Moreover, biodrainage plantations do not demand the installation of physical structures, but can be implemented by farmers themselves, who in addition may gain marketable products such as food, fodder, fiber and wood when adequate tree species are chosen. These aspects are considered to be a key element when one intends to convince farmers to introduce new technologies (Landsberg 1999a). However, the fruit species tested such as P. armeniaca and M. alba had a low potential for biodrainage purposes in the study region because of their salt sensitivity, low growth rates and resulting low DM production and WU. The tree species with largest WU and salinity tolerance, and therefore most promising for biodrainage purposes under the agro-ecological conditions in Khorezm, proved to be E. angustifolia, followed by U. pumila, P. euphratica and P. nigra var. pyramidalis. In addition to the biodrainage potential, the fruit producing and N-fixing E. angustifolia and, to a lesser extent, U. pumila also ranked top as potential suppliers of high quality supplementary fodder (Khamzina et al. 2006) and, therefore, may provide concurrent economic advantages to farmers. Conclusions Evaluating the potential of various tree species for bio-draining the saline hydromorphic soils of Khorezm should be based on field data and involve WU, salinity tolerance, rooting characteristics, and LA of the trees. Leaf transpiration rates of three-year-old trees significantly varied among species. Transpiration rates of tree species grown at sites with two soil textures and sufficiently supplied with water were highly correlated with the ability to produce and elongate the root system. Correlation of WU with 123 Agroforest Syst LA was less consistent at least in a heterogeneous young stands. The leading tree species with regards to their WU, root growth, and adaptability to the natural environment proved to be E. angustifolia followed by U. pumila, P. euphratica and P. nigra var. pyramidalis, whereas fruit species such as P. armeniaca and M. alba, though desirable from the farmer’s economic viewpoint, showed low biodrainage potential. However, since the N-fixing E. angustifolia and U. pumila have also superior feed and firewood characteristics, they may provide added value, which makes them the most suitable candidates for afforestation and biodrainage purposes. Although the species behavior during the crucial establishment phase can be considered as a prerequisite for tree further development, the ranking of species based on the evaluation of 2- to 3-year-old trees may change over time. If the observed variability in tree transpiration capacity is confirmed during subsequent years it can be used as a major criterion in selection of tree species for the biodrainage purposes. Acknowledgements The research was conducted within the framework of German/Uzbek project ‘‘Economic and ecological restructuring of land- and water use in the Region Khorezm (Uzbekistan): A pilot project in development research’’. The German Ministry for Education and Research (BMBF; project number 0339970A), the Ministry for Schools, Science and Research of the State of Northrhine-Westfalia financially supported this study. Additional funds from German Academic Exchange Service (DAAD) are gratefully acknowledged. We are thankful for the valuable comments from two anonymous reviewers on the earlier version of this manuscript. This paper is a revised, extended version of a presentation at the 21st European Regional Meeting of the International Commission on Irrigation and Drainage held in May 15– 19, 2005 in Frankfurt/Oder, Germany. References Angelocci LA, Marin FR, Oliveira RF, Righi EZ (2004) Transpiration, leaf diffusive conductance, and atmospheric water demand relationship in an irrigated acid lime orchard. Braz J Plant Physiol 16(1):53–64 Bi H, Turvey ND, Heinrich P (1992) Rooting density and tree size of Pinus radiata (D. Don) in response to competition from Eucalyptus obliqua (L’Herit). For Ecol Manage 49:31–42 123 Botman EK (1988) Improvement of sustainability and effectiveness of protective plantations in rainfed conditions in Uzbekistan. Summary of PhD thesis, Volgograd, 24 pp (in Russian) Deans JD, Munro RC (2004) Comparative water use by dryland trees in Parklands in Senegal. Agrofor Syst 60:27–38 Fimkin VP (1983) Recommendations on assortment of tree and shrub species for salt-affected soils in Golodnaya Steppe. Summary report. Middle-Asian Forestry Research Institute, Ministry of Agriculture of Uzbek SSR, Tashkent, pp 1–13 (in Russian) Forestry Compendium (2000) Global module (data base). CAB International, Wallingford, UK Hatton T, Reece P, Taylor P, McEwan K (1998) Does leaf water efficiency vary among eucalypts in water-limited environments? Tree Physiol 18:529–536 Heuperman AF, Kapoor AS, Denecke HW (2002) Biodrainage—principles, experiences and applications. International Programme for Technology and Research in Irrigation and Drainage (IPTRID) Secretariat, Food and Agriculture Organization of the United Nations (FAO), Italy, 79 pp Ibrakhimov M, Park SJ, Vlek PLG (2004) Development of groundwater salinity in a region of the lower AmuDarya River, Khorezm, Uzbekistan. In: Ryan J, Vlek PLG, Paroda R (eds) Agriculture in Central Asia: research for development. Proceedings of a symposium held at the American society of agronomy annual meetings at Indianapolis, Indiana, USA. International Center for Agricultural Research in Dry Areas (ICARDA), Aleppo, Syria, pp 56–75 Ivanov LA, Silina AA, Zelniker YuL (1950) Regarding the method of fast weighing for determination of transpiration under natural conditions. Bot J XXXV(2):1–5 (in Russian) Katz DM (1976) The influence of irrigation on ground water. Kolos, Moscow, 271 pp (in Russian) Khamzina A, Lamers JPA, Worbes M, Botman E, Vlek PLG (2006) Assessing the potential of trees for afforestation of degraded landscapes in the Aral Sea Basin of Uzbekistan. Agrofor Syst 66(2):129–141 Knight JH (1999) Root distribution and water uptake patterns in Eucalyptus and other species. In: Landsberg J (ed) The ways tree use water. Water and salinity issues in agroforestry, 5. Rural Industries Research and Development Corporation (RIRDC), pp 55–85 Kovda N, Bazilevich N, Pankova E (1985) Soil surveys for amelioration. Soviet Union collection of standards, Ministry of amelioration and water management of the USSR, Institute Soyuzgiprovodhoz, Moscow (in Russian) Landsberg J (1999a) Tree water use and its implications in relation to agroforestry systems. In: Landsberg J (ed) The ways trees use water. Water and salinity issues in agroforestry, 5. RIRDC, pp 1–27 Landsberg J (1999b) Relationship between water use efficiency and tree production. In: Landsberg J (ed) The ways tree use water. Water and salinity issues in agroforestry, 5. RIRDC, pp 45–54 Agroforest Syst Larcher W (1995) Physiological plant ecology. Ecophysiology and stress physiology of functional groups. Springer-Verlag, Berlin Heidelberg, Germany, 506 pp Martius Ch, Lamers JPA, Wehrheim P, Schoeller-Schletter A, Eshchanov R, Tupitsa A, Khamzina A, Akramkhanov A, Vlek PLG (2004) Developing sustainable land and water management for the Aral Sea Basin through an interdisciplinary research. In: Seng V, Craswell E, Fukai S (eds) Water in agriculture. Australian Centre for International Agricultural Research (ACIAR) Proceedings No 116, Canberra, pp 45–60 McMurtrie RE, Rook DA, Kelliher FM (1990) Modeling the yield of Pinus radiata on a site limited by water and nutrition. For Ecol Manage 30:381–413 Mikhailova LM (1986) Water use by fruit tree plantations in mountainous rainfed conditions (Chatkal mountain range). Protective forest planting. Proceedings of Middle Asian Forestry Research Institute, Mekhnat, Tashkent, pp 39–43 (in Russian) Mukhammadiev UK (1982) Water resource use. Tashkent, Uzbekistan (in Russian) Nnyamah JU, Black TA (1977) Rates and patterns of water uptake in a Douglas-fir forest. Soil Sci Soc Am J 41:972–979 Poljakoff-Mayber A, Lerner HR (1999) Plants in saline environments. In: Pessarakli M (ed) Handbook of plant and crop stress. Marcel Dekker Inc., New YorkBasel, pp 125–152 Santantonio D (1990) Modeling growth and production of tree roots. Process modeling of forest growth response to environmental stress. Timber Press, Portland, Oregon, pp 124–141 Schaeffer S, Willliams D, Goodrich D (2000) Transpiration of cottonwood/willow forest estimated from sap flux. Agric For Meteorol 105:257–270 Smedema B (1997) Biological drainage: myth or opportunity? GRID 11:3 Tennant D (1975) A test of a modified line intersect method of estimating root length. J Ecol 63:995–1001 Theiveyanathan V, Benyon RG (2000) Planting trees for sustainable rice production. In: Onwood; research updates from Commonwealth Scientific and Industrial Research Organization (CSIRO) Forestry and Forest Products, No 31, 5 pp 123
© Copyright 2026 Paperzz