中国科学院水利部水土保持研究所机构知识库

水分是限制干旱半干旱地区植被建设和生态恢复的最关键因素，准确获取典型植被种类蒸腾耗水特征和水分利用策略是评估生态水文效应、指导植被物种选择和优化植被配置最基础、最核心的问题。树干液流测量是研究植物整株蒸腾和植物-水分关系最有效的手段，本文（1）开发了一种适用于灌木等细茎干树干液流准确测量的外部热比率茎流计，运用数值模拟方法分析、校准树干热参数各向异性对一种五针热脉冲数字探头树干液流和热参数的测量误差并进行实测验证；（2）以黄土高原北部水蚀风蚀交错区“退耕还林（草）”背景下典型植被灌木柠条（Caragana korshinskii）、沙柳（Salix psammophila）和乔木旱柳（Salix matsudana）、小叶杨（Populus simonii）为对象，采用树干液流法监测植物蒸腾，研究植物蒸腾耗水特征和植物蒸腾对气象要素、土壤水分、土壤和地形等环境因子的响应规律。论文得到主要结果和结论如下：

（1）开发设计了一种外部热比率茎流计，经验证适用于茎干直径< 15 mm和< 50 cm³ cm^-2 h^-1液流密度准确测量。人工导水-称重法校准试验表明，茎流计测得热脉冲速率（V_h）信号和实际液流密度（V_s）极显著线性相关(R² = 0.95，P < 0.001），但是柠条和沙柳由V_h转换成V_s的校准系数m_sap具有显著差异，差异来源于两种灌木的茎干结构差异，建议实际运用时对不同类型茎干分别校准。野外测试表明，外部热比率茎流计能够准确测定低速和逆向液流。分析模拟树干木质部热参数各向异性对基于INV-WATFLX算法的五针热脉冲数字探头树干液流和热参数的测量误差，结果显示，不同探头安装角度α下各向异性对液流密度的测量误差-30% ~ +40%，零液流条件下热参数（热扩散率κ、热导率λ和热容量C）测量误差 < 12%；本研究提出了相应校准模型。野外实测显示，探头以15°和30°安装时，可以准确获取树干液流和热参数；但是以0°安装时树干液流低估约30%，可能原因是以0°安装时探头其中共平面的三针的插入，阻碍了液流在轴向液流路径中的传输，导致液流低估。但零液流条件下，探头以0°安装时可以获得更稳定的热参数测量结果。

（2）黄土高原北部水蚀风蚀交错区7 a龄沙柳枝干液流密度和整株蒸腾量均显著高于15 a龄柠条。在2017年7-9月，沙柳平均液流密度8.26 cm³ cm^-2 h^-1，平均整株蒸腾2.6 kg d^-1 （或2.22 mm d^-1）；柠条平均液流密度5.43 cm³ cm^-2 h^-1，平均整株蒸腾1.92 kg d^-1（或0.77 mm d^-1）。气象条件（参考蒸散ET₀、太阳辐射R_n和饱和水汽压亏缺VPD）是驱动柠条和沙柳蒸腾的最重要环境因素，和蒸腾强度具有极显著正相关关系（P < 0.01）。对于柠条，单因素ET₀可解释日蒸腾变异91%，R_n可解释日蒸腾变异83%，VPD可解释蒸腾变异77%；对于沙柳，单因素ET₀可解释蒸腾变异77%，R_n可解释蒸腾变异79%，VPD可解释蒸腾变异61%。偏相关分析表明，柠条蒸腾和0-100 cm深度土壤含水量无显著相关关系（P > 0.05），而沙柳蒸腾和0 – 50 cm深度土壤含水量显著相关（P < 0.05）。根系调查显示，沙柳水平根发达，根系聚集在土壤浅表层，以快速高效吸收当季降水转化的浅层土壤水；柠条细根在浅表土层分布较少，细根根长密度仅是沙柳的1/2，但具有发达的垂直深根系，可持续利用深层土壤水分，而避免遭受浅层土壤干旱胁迫的影响。柠条是节水型耐旱灌木，深层土壤水是重要水分利用来源；沙柳是耗水型植被，主要吸收利用浅层土壤水。鉴于柠条比沙柳更加节水耐旱和具备更强的抵抗根-土分离能力，因此建议柠条更适用于水蚀风蚀交错区的植被恢复。

（3）通过对水蚀风蚀交错区风沙土坡地和黄绵土坝地约30 a龄人工种植旱柳和小叶杨最长达连续7年的生长季树干液流、气象要素和土壤水分等同步监测研究，结果表明，坝地旱柳和小叶杨生长季平均整株蒸腾分别是46.7和175.2 kg d^-1，分别是坡地相同树种的5.6和4.2倍；整体而言小叶杨整株蒸腾耗水量是旱柳约4倍。气象要素（ET₀、R_n和VPD）是影响旱柳和小叶杨最重要的环境因子，均呈极显著正相关关系（P < 0.01）。ET₀可解释坡地旱柳蒸腾变异37%，坝地旱柳蒸腾高达77%，坡地和坝地小叶杨蒸腾变异80%。坡地旱柳蒸腾与0-200 cm深度浅层各层土壤含水量显著正相关（P < 0.05）；但对坡地小叶杨和坝地旱柳和小叶杨，蒸腾和浅层土壤含水量没有一致的相关关系。丰水年降水补给风沙土地土壤水分深度达600 cm，而在干旱年仅120 cm；在平水年和干旱年，风沙土地0–600 cm产生土壤水分亏缺，在丰水年得到缓解；浅层和深层土壤水分均是风沙土地树木重要水分来源。而对于黄绵土地，即使丰水年，补给深度也仅200 cm，坝地旱柳和小叶杨主要消耗浅层地下水。坡地树木细根在浅层土壤聚集，而在坝地无此现象，是旱柳和小叶杨根系对区域降水规律和水分利用策略的适应性。从水资源可持续性角度考虑，两种高蒸腾耗水树种均不适合在黄土高原水蚀风蚀交错区广泛种植。

Water availability is the key restricting factor for vegetation construction and ecological restoration in arid and semi-arid area. Accurately quantifying the revegetated plant transpiration characteristics and understanding their water use strategies is essential to evaluate their ecohydrological effects, guide revegetated plant species selection and optimize revegetated plant composition configuration. This dissertation (1) designed an external heat-ratio sap flow (EHR) gauge suitable for thin stem plants as shrubs accurate sap flow measurements, and numerically simulated the effects of isotropy in sapwood thermal properties on sapwood sap-flux density and thermal properties sensing using a smart penta-needle heat-pulse probe (PHPP), developed a calibration model and validated in field tests; (2) investigated plant transpiration characteristics using sap flow methodologies and its responses to meteorology, soil water conditions, soil texture and topography, to typical revegetated shrub species Caragana korshinskii (C. k.) and Salix psammophila (S. p.) and tree species Salix matsudana (S. m.) and Populus simonii (P. s.) in the water-wind erosion crisscross region on the Northern Loess Plateau, under the background of Grain-for-Green project. This thesis mainly summarized and concluded as follows:

(1) We developed a new EHR sap flow gauge, validated suitable for accurate sap-flux density (V_s) measurement of thin stem with diameter < 15 mm and with V_s < 50 cm³ cm^-2 h^-1. Results of artificial induced flow-gravimetric method test indicated that, the output heat pulse velocity (V_h) signal by EHR gauge was significantly linear correlated with actual V_s (R² = 0.95，P < 0.001), but the coefficients, m_sap, transferring V_h into V_s were apparently different, and the potential cause was the difference in stem physical structure between C. k. and S. p., we strongly recommended separate gauge calibration experiment for different plant species. Field tests indicated that, EHR gauge is capable for accurate measurement of low rate and reverse flow in plant stems. We numerically simulated the effects of isotropy in sapwood thermal properties on sapwood V_s and thermal properties sensing using a smart PHPP coupled with INV-WATFLX algorithm, results indicated that, under different probe installation angle α, errors in V_s measurement induced by isotropicy were ranged from -30% ~ +40%, and errors in thermal properties (thermal diffusivity, κ, thermal conductivity, λ, and heat capacity, C) under zero-flow conditions were within 12%; we developed a calibration model. Field tests indicated that, when PHPPs installed at α = 15° and 30°, V_s and thermal properties were accurately obtained; but at α = 0°, obtained V_s were underestimated by 30%, and the potential cause is the flow obstruction by insertion of three needles in a plane parallel to stem axial direction. But under zero-flow conditions, PHPP installed at α = 0° would obtain more steady thermal properties.

(2) In study plots at the water-wind erosion crisscross region on the Northern Loess Plateau, results indicated that, mean V_s and whole-pant transpiration (T) of the 7–year-old S. p.stems were significant higher than that of C. k. During July to September, 2017, mean V_s of S. p. was 8.26 cm³ cm^-2 h^-1, and mean whole-plant T was 2.6 kg d^-1 (or 2.22 mm d^-1); mean V_s of C. k. was 5.43 cm³ cm^-2 h^-1, and mean whole-plant T was 1.92 kg d^-1 (or 0.77 mm d^-1). Meteorological factors (reference evapotranspiration, ET₀, solar radiation, R_n, and vapor-deficit pressure, VPD) is the most important driving factor of C. k. and S. p., and they were significantly positively linear correlated with T (P < 0.01). For C. k., univariate ET₀ interpreted 91% of the daily whole-plant T variation, univariate R_n interpreted 83%, and univariate R_n interpreted 83%; for S. p., univariate ET₀ interpreted 77% of the daily whole-plant T variation, univariate R_n interpreted 79%, and univariate R_n interpreted 61%. Partial correlation analysis indicated that, there were not significant correlationship between T of C. k. and soil water contents (SWC) in soil depth 0 – 100 cm; but T of S. p. were significantly correlated with SWC in soil depth 0 – 50 cm. Root investigation indicated that, S. p. had well-developed lateral root, and lateral root gathered in shallow soil layers to rapidly and effectively absorb shallow soil water transferred from rainfall; C. k. had fewer lateral root in shallow soil layers, mean fine-root-length-density (RLD) was only half that of S. p., but it had well-developed vertical deep-root system, which could steadily absorb deep soil water, to avoid being affected by suffering drought stress in shallow soil layers. C. k uses water-saving and drought tolerance water-use strategy, and it uses deep soil water as a substantial water sources; S. p. uses water-consumption water-use strategy, and it uses shallow soil water as main water resources. Given the fact that C. k. consumes less soil water, has higher drought tolerance and soil reinforcement intermedia root and soil, we recommended C. k. prefer to S. p. ­to use in revegetation practices in wind-water erosion crisscross region.

 (3) By long-term (at most up to seven years) and simultaneous monitoring of sap flow, meteorological factors, soil water conditions and et al. for > 30-year-old S. m. and P. s. in an aeolian sandy soil sloping field and a loessial soil dam field in the water-wind erosion crisscross region, results indicated that, mean whole-tree T of S. m. and P. s. in dam field in growing season were 46.7 and 175.2 kg d^-1, respectively, and they were 5.6 and 4.2-fold of the same species in sloping field, respectively; overall, whole-tree T of P. s. was about 4-fold of S. m. Meteorological factor (ET₀, R_n and VPD) were the controlling factors driving tree T of S. m. and P. s., and they were significantly correlated with tree T (P < 0.01). Univariate ET₀ interpreted 37% of variation in S. m. T in sloping field, 77% for S. m. in dam field, and 80% for P. s. in both sloping field and dam field. T of S. m. in sloping field were significantly correlated with shallow layer SWCs in 0 – 200 cm depth (P < 0.05); but for P. s. in sloping field and S. m. and P. s. in dam field, there had no consistent correlationship between T and SWCs in shallow soil layers. Annual rainfall soil-water recharge depth reached up to 600 cm in wet years, but only 120 cm in dry years on sloping field with aeolian sandy soil; soil water deficit formed in the 0 - 600 cm soil layer on the aeolian sandy soil field in normal and dry years, and soil water was replenished during wet years; both shallow and deep soil water are primary water sources for trees in aeolian sandy soil sloping field. For loessial soil, recharge depth was only 200 cm even in wet years in dam field, and trees on the dam field absorbed water mainly from shallow groundwater. Tree fine roots gathered in the subsurface soil on the sloping field but not in the dam field indicated a root distribution adaption to precipitation patterns and water-use strategy. Neither species is suitable for extensive revegetation implementation in this area due to their negative impacts on water resource sustainability.