We studied the effects of three nitrogen (N) supply rates (low, intermediate and high) on Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) seedlings and poplar clone “I-214” (Populus x euroamericana (Dole) Guinier) cuttings growing in mini-stands. Our specific objectives were to: (1) evaluate the effects of N supply on water-use efficiency (WUE) and biomass production; (2) determine if N affects WUE through control of carbon assimilation rates or through stomatal control of water loss; and (3) compare three methods of estimating WUE: one short-term method (WUE_i, based on gas exchange measurements) and two long-term methods (WUE_T, based on the ratio between biomass production and transpired water, and Δ, based on leaf carbon isotope discrimination tested as a proxy of WUE). In both species, biomass production, WUEi and WUE_T increased with increasing N supply, but there was no effect of N supply on either transpiration or stomatal conductance and Δ was negatively related to leaf N concentration. Plots of Δ versus both WUE_i and WUE_T revealed negative trends, but the regression between WUE_i and Δ was significant only for Douglas-fir, and the regression between WUE_T and Δ was significant only for poplar. Thus, the mechanisms underlying the response of WUE to N supply were mainly related to a positive effect of N supply on photosynthetic rates. The data confirm that carbon isotope discrimination may be a useful proxy of WUE. The finding that N availability enhances both biomass production and WUE may have practical implications in regions where these factors impose constraints on forest productivity.

Monitoring soil water content at high spatio-temporal resolution and coupled to other sensor data is crucial for applications oriented towards water sustainability in agriculture, such as precision irrigation or phenotyping root traits for drought tolerance. The cost of instrumentation, however, limits measurement frequency and number of sensors. The objective of this work was to design a low cost “open hardware” platform for multi-sensor measurements including water content at different depths, air and soil temperatures. The system is based on an open-source ARDUINO microcontroller-board, programmed in a simple integrated development environment (IDE). Low cost high-frequency dielectric probes were used in the platform and lab tested on three non-saline soils (ECe1: 2.5 < 0.1 mS/cm). Empirical calibration curves were subjected to cross-validation (leave-one-out method), and normalized root mean square error (NRMSE) were respectively 0.09 for the overall model, 0.09 for the sandy soil, 0.07 for the clay loam and 0.08 for the sandy loam. The overall model (pooled soil data) fitted the data very well (R² = 0.89) showing a high stability, being able to generate very similar RMSEs during training and validation (RMSE_training = 2.63; RMSE_validation = 2.61). Data recorded on the card were automatically sent to a remote server allowing repeated field-data quality checks. This work provides a framework for the replication and upgrading of a customized low cost platform, consistent with the open source approach whereby sharing information on equipment design and software facilitates the adoption and continuous improvement of existing technologies.

Traditional methods for studying tree roots are destructive and labor intensive, but available nondestructive techniques are applicable only to small scale studies or are strongly limited by soil conditions and root size. Soil electrical resistivity measured by geoelectrical methods has the potential to detect belowground plant structures, but quantitative relationships of these measurements with root traits have not been assessed. We tested the ability of two-dimensional (2-D) DC resistivity tomography to detect the spatial variability of roots and to quantify their biomass in a tree stand. A high-resolution resistivity tomogram was generated along a 11.75 m transect under an Alnus glutinosa (L.) Gaertn. stand based on an alpha-Wenner configuration with 48 electrodes spaced 0.25 m apart. Data were processed by a 2-D finite-element inversion algorithm, and corrected for soil temperature. Data acquisition, inversion and imaging were completed in the field within 60 min. Root dry mass per unit soil volume (root mass density, RMD) was measured destructively on soil samples collected to a depth of 1.05 m. Soil sand, silt, clay and organic matter contents, electrical conductivity, water content and pH were measured on a subset of samples. The spatial pattern of soil resistivity closely matched the spatial distribution of RMD. Multiple linear regression showed that only RMD and soil water content were related to soil resistivity along the transect. Regression analysis of RMD against soil resistivity revealed a highly significant logistic relationship (n = 97), which was confirmed on a separate dataset (n = 67), showing that soil resistivity was quantitatively related to belowground tree root biomass. This relationship provides a basis for developing quick nondestructive methods for detecting root distribution and quantifying root biomass, as well as for optimizing sampling strategies for studying root-driven phenomena.