Design, construction, and calibration of an automated weighing lysimeter to support climate-smart agriculture in the Colombian Orinoquía

Abstract

In tropical agricultural regions, accurate estimation of reference evapotranspiration (ET₀) remains challenging due to climatic variability, data scarcity, and limitations of empirical models such as FAO-56 Penman–Monteith. To provide the technical basis for future field-based evapotranspiration studies, this study reports the design, construction, and calibration of an automated weighing lysimeter installed at the La Libertad Research Center in Villavicencio, Colombia. The system integrates a galvanized steel tank (1 m³) reconstructed with original soil horizons, an automated sprinkler irrigation system, and a monitoring module based on load cells, soil moisture sensors, a professional weather station, and real-time data transmission powered by a photovoltaic unit. Calibration was performed using incremental load–unload cycles, achieving linear response curves with R² > 0.99, while soil water retention curves were developed to define irrigation thresholds. The results document the successful construction and integration of the system and show highly linear load-cell responses, also the capacity to provide high-resolution measurements under equatorial conditions, this could reduce methodological uncertainties inherent to model-based approaches. The developed system provides the technical infrastructure required for future evapotranspiration monitoring and comparison with established ET₀ estimation methods. Overall, the proposed lysimeter may support future irrigation-management, water-resource assessment, and climate-smart agriculture applications in tropical regions.

1 Introduction

In Colombia, the energy, domestic, and agricultural sectors have the highest water demands, but estimating this demand is challenging due to climate variability, geographical diversity, and limited data availability. In 2020, the Colombian Institute of Hydrology, Meteorology, and Environmental Studies (IDEAM) reported a multisectoral water demand of approximately 33 million m³, with nearly 14 million m³ used by agriculture, highlighting its key role in national food security. The Department of Meta, selected as the study area, is notable for high water demand, accounting for 6.38% of the national total, with 66% of this used for irrigation and post-harvest processes (Instituto de Hidrología and Meteorología y Estudios Ambientales IDEAM, 2023).

It is essential to note that one key indicator of water resource management is the so-called blue water footprint, which refers to the volume of surface and groundwater utilized for irrigation. Another indicator, the green water footprint, measures water uses in rainfed agriculture systems and refers to sources such as rainfall and runoff (Instituto de Hidrología and Meteorología y Estudios Ambientales IDEAM, 2023). To determine and estimate both indicators, empirical and physico-mathematical methods are commonly applied in Colombia, with the Penman-Monteith model (PMM) being the most widely used (Gómez Blanco and Cadena, 2018; Calderón Castaño, 2017; Allen et al., 2006). However, this model still has flaws associated with the energy balance closure, which is probably related to the overestimation of the aerodynamic component of ETo (Hao et al., 2018).

In this way, despite the widespread application of the model, its outputs exhibit significant uncertainty, particularly in Equatorial Regions, which have the unique climatic conditions described by Henry (Henry and Oliver, 2005). Furthermore, the PMM requires multiple climatic variables as input, but such data are not always accessible, complete, or of adequate quality to support accurate estimations (Calderón Castaño, 2017). This presents a distinct challenge in remote agricultural areas, such as the Colombian Eastern Plains, a vast region experiencing rapid agrarian transformation (Devia Acosta and Piñeros Lizarazo, 2021), where limited access hinders the installation of weather stations. As a result, irrigation requirements for infrastructure design may be either overestimated or underestimated.

Some authors (Gómez Blanco and Cadena, 2018; Calderón Castaño, 2017; Mendoza Castiblanco and Peña Quiñones, 2021; Jaramillo Robledo, 2006) have compared the performance of equations or models used to calculate reference evapotranspiration (ETo) in Colombia. However, no studies have compared the output data from such empirical or physical-mathematical models with measured evapotranspiration using more precise methods, such as lysimetry. This highlights the need to evaluate the performance of the Penman–Monteith approach for estimating reference evapotranspiration under the specific climatic conditions of the Colombian Orinoquía. Therefore, there is a need to conduct accurate in situ measurements using devices such as weighing lysimeters. These devices can provide real-time data and reduce uncertainties associated with empirical models (Allen et al., 2006; Arruda Junior and da Silva, 2023; Carvalho et al., 2007). By measuring mass changes within an isolated soil profile, weighing lysimeters provide direct water-balance observations that can be used as a reference for evaluating model-based evapotranspiration estimates (Arruda Junior and da Silva, 2023; Carvalho et al., 2007; Howell et al., 1991).

The primary objective of this study was to design, construct, instrument, calibrate, and perform an initial field verification of an automated weighing lysimeter in Villavicencio, Meta. The developed system integrates load-cell weighing, soil-moisture monitoring, automated irrigation, meteorological measurements, autonomous power supply, and Internet of Things (IoT)-based data transmission. This study focuses on the engineering implementation of the lysimeter, the characterization of the reconstructed soil profile through soil water retention curves, and the calibration response of the weighing components. Long-term evapotranspiration measurements and comparisons with established estimation methods, including FAO-56 Penman–Monteith (Allen et al., 2006), will be addressed in subsequent stages of the research. In this paper, the authors report on how recent advances in low-cost technology and the Internet of Things (IoT) can improve the collection of high-precision evapotranspiration data (Arruda Junior et al., 2023). The results obtained will be essential for optimizing irrigation management in regional crops, and they may also influence current national water footprint estimations. Furthermore, this study aims to establish a foundation for future research focused on developing evapotranspiration models adapted to Colombia’s specific climatic and geographic conditions.

2 Materials and methods

The lysimeter’s design takes into account its deployment environment. It will be installed in the Colombian Orinoquía, in a transitional landscape from Piedmont (piedemonte, the foothills at the base of the Andes) to High Plains (altillanura, extensive flat upland plains). The site is La Libertad Research Center, part of the Colombian Agricultural Research Corporation (AGROSAVÍA) (Figure 1). Coordinates are 4.057°N, −73.467°W, at 336 m above sea level. The average annual precipitation is 2,933 mm distributed in two seasons (rainy and dry), the rainy one from April to November, when falls around 90% of the annual total, and the dry season between December and March (Corporación Colombiana de InvestigaciónAgropecuaria, 2025). The average maximum monthly air temperature is higher during the dry season, while the average minimum air temperature does not exhibit significant variations throughout the year. The lysimeter was installed in an open field, away from vegetation, to ensure representative climatic exposure and minimize boundary effects.

FIGURE 1

2.1 Lysimeter design and construction

The design and construction of this weighing lysimeter were based on the infrastructure proposed by (Allen et al., 2006; Arruda Junior and da Silva, 2023; Carvalho et al., 2007; Dong and Hansen, 2023; Gomes Junior et al., 2019; Nascimento et al., 2011). For this study, we modified the container dimensions and adjusted the number and arrangement of load cells to accommodate local environmental conditions, enhance accuracy, and meet the specific needs of the chosen herbaceous crop species. These changes were made to meet the region’s requirements and the needs of the selected crops (herbaceous species). Table 1 lists the materials and equipment used in this study. Figure 2 illustrates the methodological process followed for the design, construction, and implementation of the proposed lysimeter.

FIGURE 2

2.2 Lysimeter main components

In the conceptual design of the lysimeter, four main components were identified and characterized for their construction. First is the Lysimeter Isolation and Protection System (LIPS), illustrated in Figure 3. Next is the metallic Lysimeter Tank (LT), shown in Figure 4. The third component is the Irrigation System (IS), presented in Figure 5. Finally, the Monitoring, Measurement, Data Storage, Transmission, and Power Supply System (MMDSTP) is detailed in Figure 6.

FIGURE 3

FIGURE 4

FIGURE 5

FIGURE 6

2.2.1 The Isolation and Protection System (LIPS)

It consists of a masonry structure designed to isolate and support the soil surrounding the lysimeter tank. By doing so, this design enables the exclusive measurement of the soil matrix contained within the lysimeter, while maintaining the original conditions of the surrounding soil and vegetation. As illustrated in Figure 3, the dimensions of this structure are defined by concrete columns and beams. Furthermore, the four masonry walls were built using bricks and have an internal coating of cement and sand, with an average thickness of 2 cm.

2.2.2 Lysimeter Tank (LT)

The tank is specifically designed to contain both the soil matrix and the crop to be cultivated within it (Figure 4). This setup provides a controlled environment for measuring variables such as soil moisture and changes in the total weight of the matrix and the crop. To achieve this, strategic reinforcements were incorporated into the tank’s walls and base to support the volume and ensure structural integrity over time (Figure 4b). As a result, the galvanized steel tank can hold 1 m³ of soil.

2.2.3 Irrigation System (IS)

The system is designed to supply the water needed to keep the soil matrix consistently near its field capacity. Its design is based on a prior analysis of the water requirements for soil and crops. This analysis examines evapotranspiration rates and the soil’s water retention capacity. The hydraulic analysis ensures the sprinkler irrigation system provides uniform distribution across the entire soil profile and application area (Figure 5).

2.2.4 Monitoring, measurement, data storage, transmission, and power supply system (MMDSTP)

The system is designed to efficiently capture, process, store, and transmit information from its sensors and devices. The technical criteria and key parameters for selecting each component are detailed in Figures 6, 7.

FIGURE 7

2.3 Calibration and validation procedures

To accurately measure water lost due to crop surface evapotranspiration, the main component of the MMDSTP, the load cell, was calibrated. This process included two main stages. First, soil water retention curves (SWRC) were obtained to determine field capacity and permanent wilting point values within the soil matrix. Second, load and unload curves for the load cells were generated. Finally, in situ measurements were performed using known weights with the entire system fully installed and operational. For this in situ verification, three filled 5-gallon water containers were sequentially placed on the lysimeter as controlled incremental loads. The corresponding stepwise changes recorded by the integrated weighing and data-acquisition system were monitored to verify the response of the complete assembly, including the tank, reconstructed soil profile, load cells, and datalogger, under its actual field configuration. This procedure was conducted as an initial functional check before continuous field monitoring began.

2.3.1 Soil sampling

To determine the SWRC for the soil used in the lysimeter, the first step involved excavating a soil volume of 1 m³ with dimensions of 1 m × 1 m × 1m (Figure 8a). During this process, soil samples were collected to develop the water retention curves. The three identified horizons (Ap, AB, and Bw) were carefully separated. Each sample was labeled and stored in sealed containers next to the excavation area (Figure 8b). The soil samples were submitted to Richard´s pressure plate extractors. This method measures the link between soil water content and suction (Guillaume et al., 2023). Measurements were made at the Soil Physics Laboratory of the National University of Colombia, Bogotá. Eighteen samples were taken using steel rings (0.8” high, 2” diameter), divided equally among the three horizons.

FIGURE 8

These horizons were classified at the sampling site as Ap (plowed or cultivated horizon), AB (transition horizon), and Bw (incipient or developing horizon) following the USDA soil-classification guidelines (Natural Resources Conservation Service, 2022). The samples were placed in two pressure extractors operating in parallel. Each extractor held three samples per horizon (Figure 8c). The samples were subjected to 11 pressure levels ranging from 0.03 to 1.4 MPa. The upper pressure limit was determined by the maximum operating capacity of the compressor connected to the pressure-plate extraction system. Consequently, direct measurements at the conventional pressure of 1.5 MPa could not be performed. To improve the characterization of the high-suction portion of the soil water retention curve, additional intermediate pressure levels were included near the upper operating range of the equipment. These measurements provided a denser experimental dataset to support the subsequent estimation of volumetric water content at 1.5 MPa. The IGAC analytical procedure considers ceramic pressure-plate extractors operating within a pressure range of 0.03–1.5 MPa (Instituto Geográfico Agustín Codazzi IGAC, 2006).

For each soil horizon, a local linear regression was subsequently fitted to the five measurements obtained within the upper-pressure range of 1.0–1.4 MPa. The fitted relationships were used to extrapolate volumetric water content over the limited interval from 1.4 to 1.5 MPa and estimate the permanent wilting point at the conventional pressure of 1.5 MPa.

2.3.2 Load cells calibration

For the calibration of the load cells, load and unload curves were obtained. This procedure was carried out in a soil laboratory using an unconfined compression testing press. The press had been calibrated under ISO/IEC 17025:2017 by an accredited laboratory (ICCLAB). A functionally equivalent device to the 50 kN Unconfined Compression Tester developed by ELE International was used in this study (Figure 9). Each load cell was initially loaded with 10 kg. The load was then increased in 50 kg increments up to 450 kg. Beyond this point, the load was increased in 2 kg increments until it reached 500 kg. This provided greater accuracy under conditions like those of the lysimeter. The unloading process followed the same methodology in reverse. It started at 500 kg and decreased in 2 kg steps down to 450 kg, and then in 50 kg intervals down to 10 kg. This calibration procedure was adapted from the load-cell calibration protocol developed by the Ferguson Structural Engineering Laboratory (FSEL) at The University of Texas at Austin (Ferguson Structural Engineering Laboratory FSEL, 2016). The protocol describes controlled loading and unloading cycles performed with a uniaxial testing machine and the comparison of the sensor response with a calibrated reference system.

FIGURE 9

Two complete loading and unloading cycles were performed for each load cell under the same controlled laboratory conditions. Separate linear calibration equations were derived for each load cell under loading and unloading conditions. During lysimeter operation, the signal from each cell was converted to load using its corresponding individual calibration equation, and the four calculated loads were subsequently summed to obtain the total lysimeter mass.

2.3.3 Software and data processing

Load-cell calibration data and soil water retention measurements were organized and processed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). This software was used to calculate the individual linear regressions, regression coefficients, coefficients of determination, and the consolidated calibration relationships. Microsoft Excel was also used to fit the local linear regressions applied to the high-pressure portion of the soil water retention curves, estimate volumetric water content at 1.5 MPa, and prepare the graphical representations.

The CR1000X datalogger program was developed using CRBasic Editor, while LoggerNet software (Campbell Scientific, Inc., Logan, UT, USA) was used to upload the program to the datalogger, establish communication with the system, configure sensor connections and communication parameters, and retrieve stored measurements remotely when required. Three-dimensional models and engineering drawings of the lysimeter components and integrated system were developed using Autodesk AutoCAD (Autodesk, Inc., San Francisco, CA, United States).

3 Results and discussion

3.1 Structural assembly and implementation of the lysimeter system

The construction phase of the LIPS followed the design protocol specifications. Initial work involved masonry, including the casting of reinforced concrete beams. These beams supported the structural frame. They also served as foundations for four brick walls. The walls, built using fired clay units and a 1 cm cement-sand mortar layer, enclosed the system (Figure 10).

FIGURE 10

Within the same area, four precast concrete blocks (0.008 m³ each) were placed and anchored as shown in (Figure 11a). These provided firm support for the weighing tank, keeping it properly aligned and strong under both static and dynamic loads. The LT was built using 14-gauge galvanized steel (2 mm thick), chosen for its strength and resistance to rust. The tank was shaped by cutting and joined together by MIG welding (Figure 11b). Additional support was provided by a mesh of 1-inch galvanized steel pipes spaced along the tank walls to bear soil pressure (Figure 11b). Three H-shaped bars (1/16” thick) at the base help carry weight and allow moving the tank with a forklift (Figure 11c).

FIGURE 11

Before adding soil, we filled the tank with water to check for leaks along the welds. We addressed any issues to ensure the tank was watertight. The complete structure met the strength and leak-proof standards set in the engineering design (Figure 11d).

3.2 Internal configuration and soil profile restoration

The base of the lysimeter tank was prepared with a 10 cm gravel layer (6–8 mm particles), conforming to Colombian standard NTC 174 (Ins tituto Colombiano de Normas Técnicas y CertificaciónICONTEC, 2000), to promote internal drainage (Figure 12a). A perforated 1½” PVC pipe was embedded within the gravel to collect and direct percolated water away from the system (see Figure 12a). Above this, a nonwoven polypropylene geotextile (400 g/m²) was placed, cut to match the internal footprint of the tank. The geotextile provided mechanical separation and filtration in accordance with the technical criteria for geotextiles used in filtration and drainage applications established in EMCALI Technical Standard NDC-SE-GE-003/V3.0 (Empresas Municipales de Cali EMCALI EICE ESP, 2026), preventing soil particle migration while maintaining drainage continuity (Figure 12b). The metallic tank and the surrounding LIPS structure prevent uncontrolled lateral water exchange between the reconstructed soil profile and the surrounding ground, thereby defining an isolated soil-water control volume. The gravel layer and perforated drainage pipe provide a free-draining lower boundary, reducing the risk of artificial water accumulation and perched saturation at the bottom of the lysimeter. Water losses through drainage produce a corresponding decrease in the total mass recorded by the load-cell system, while changes in water storage within the soil profile are independently monitored by the soil-moisture sensors. The reconstructed horizons and exposed soil surface allow the lysimeter to experience atmospheric conditions comparable to those of the surrounding field. Therefore, the drainage configuration was intended to reproduce representative soil-water dynamics rather than to assume identical moisture content inside and outside the lysimeter at every moment.

FIGURE 12

The soil profile was reconstructed in stratified layers using previously excavated, sieved, and homogenized materials, replicating the original pedological horizons. Filling was performed incrementally to minimize compaction variability and to maintain the natural bulk density of the soil (Figures 12c,d).

3.3 Irrigation system integration

To maintain the soil matrix near field capacity, an automated irrigation system was designed and installed. Water conveyance was achieved through a ½” PVC pipeline buried along a 100-m trench with a 2% slope and 40 cm depth to ensure physical protection and gravitational assistance (Figure 13a). Pipe joints were sealed using primer and solvent cement. The delivery system consisted of a 12 V DC solenoid valve, a diaphragm pump (2.8–3.0 GPM, 45 PSI), and a rotary sprinkler with a coverage radius of 6–10 m and a flow rate of 100–200 L/h (see Figure 13b for system components). Component selection was based on hydraulic compatibility and durability under experimental field conditions.

FIGURE 13

3.4 Monitoring, data acquisition, and power system (MMDSTP)

The lysimeter’s monitoring and automation system was built around a CR1000X datalogger (Campbell Scientific, 2026), programmed to process and transmit data from the load cells and TEROS 12 soil sensors (METER Group, 2024). These sensors measure volumetric water content (VWC), temperature, and electrical conductivity using the SDI-12 protocol.

Meteorological variables, including air temperature, relative humidity, wind speed, solar radiation, and precipitation, were captured using a professional-grade weather station (Gill Instruments, 2022) integrated into the system. All data were transmitted in real time via a Sierra Wireless modem connected to an omnidirectional antenna.

Power was supplied through a 100 W photovoltaic panel, a charge controller, and a 12 V, 100 Ah AGM GEL battery, ensuring uninterrupted operation under autonomous conditions (Figure 14a). The complete electrical installation was housed in a 60 × 40 × 23 cm polyester enclosure with internal rail guides, cable management, and surge protection (Figure 14b).

FIGURE 14

3.5 Soil water retention curve and irrigation thresholds

The soil water retention curves shown in Figure 15 describe the average volumetric water content measured for the Ap, AB, and Bw horizons over the experimental pressure range of 0.03–1.4 MPa. Because the maximum operating pressure of the available extraction system was 1.4 MPa, permanent wilting point values at the conventional pressure of 1.5 MPa were not measured directly. Instead, local linear regressions were fitted to the five measurements obtained between 1.0 and 1.4 MPa for each horizon and extrapolated over the limited interval to 1.5 MPa. The resulting estimates were 22.55% VWC for Ap, 24.77% for AB, and 26.61% for Bw. The field-capacity reference values measured at 0.03 MPa and the estimated permanent-wilting-point values at 1.5 MPa are reported in Table 2, while the original pressure-plate data are available in the Mendeley Data repository (Caceres, 2025).

FIGURE 15

3.6 Load cell calibration

Two complete loading and unloading cycles were performed for each of the four load cells across the full operational range under controlled laboratory conditions. Figures 16a-h presents one representative loading and unloading cycle for each sensor. The calibration data yielded highly linear response curves, with coefficients of determination greater than (R² > 0.995). These coefficients demonstrate the linearity of the sensor responses within the evaluated load range but should not be interpreted as a formal assessment of long-term stability or measurement uncertainty.

FIGURE 16

Separate linear calibration equations were obtained for each load cell under loading and unloading conditions. During lysimeter operation, the voltage output of each sensor was converted to load using its corresponding individual calibration equation. The four calculated load values were then summed to determine the total lysimeter mass. Therefore, a single common calibration equation was not applied to all four sensors. For descriptive purposes, the four individual calibration relationships were algebraically combined to obtain consolidated equations representing the overall response of the weighing system under loading and unloading conditions:

Where x is the applied load (kg) and y is the corresponding voltage output (mV). These consolidated equations summarize the system-level response and were not used to replace the individual calibration equations applied to each load cell. Figures 16a-h shows the representative individual loading and unloading calibration curves and their corresponding linear regressions. The datasets supporting the calibration curves presented in the manuscript are available in the Mendeley Data repository (Caceres, 2025).

The construction and deployment of the weighing lysimeter demonstrate the feasibility of integrating structural, hydraulic, sensing, telemetry, and autonomous power components under field conditions in the Colombian Orinoquía. Arachis pintoi cv. Centauro was established as permanent ground cover for subsequent monitoring; however, crop evapotranspiration performance was not evaluated in the present study.

The use of 2 kg load increments between 450 and 500 kg increased the density of calibration points near the upper operational range of the load cells and allowed a more detailed characterization of their response under conditions similar to those expected during lysimeter operation. This procedure improved the resolution of the calibration curves; however, hysteresis was not quantified independently, and therefore no claim is made that the selected load increments reduced hysteresis. Previous weighing-lysimeter studies have similarly emphasized the importance of calibration across the operational load range (Arruda Junior and da Silva, 2023; Nascimento et al., 2011). The individual calibration regressions obtained in this study showed coefficients of determination greater than 0.99, confirming a highly linear response within the evaluated range.

Environmental disturbances, including vibration and temperature fluctuations, can affect weighing-lysimeter measurements (Dong and Hansen, 2023). Accordingly, the LIPS and MMDSTP systems were configured to support autonomous operation under humid equatorial field conditions. The photovoltaic panel and battery reduce dependence on grid electricity, while the cellular modem enables remote data transmission. In addition, the soil water retention analysis provided operational soil-moisture reference values for irrigation control within the reconstructed soil profile. Together, these components establish the technical basis for future studies of evapotranspiration, irrigation management, and agricultural water use in the Colombian Orinoquía (Instituto de Hidrología and Meteorología y Estudios Ambientales IDEAM, 2023; Allen et al., 2006).

Furthermore, the developed lysimeter is intended to support future evaluation and, where necessary, local adjustment of widely used reference evapotranspiration estimation methods, including FAO-56 Penman–Monteith, and of crop coefficients used to derive actual crop evapotranspiration. Previous Colombian studies have compared alternative methods for estimating reference evapotranspiration (Gómez Blanco and Cadena, 2018; Calderón Castaño, 2017; Mendoza Castiblanco and Peña Quiñones, 2021), while studies conducted in other environments have directly compared model-based estimates with lysimeter measurements (Santiago, 2002; Kandra et al., 2023). These comparisons were not performed in the present study and will require analysis of the long-term field dataset. Because model performance depends on the availability and quality of meteorological inputs and may vary among climatic and surface conditions (Hao et al., 2018), future lysimeter observations will provide an independent field reference for assessing the applicability of these methods under the humid equatorial conditions of the Colombian Orinoquía.

Additionally, the developed system is consistent with recent advances in low-cost and digitally connected weighing lysimetry. Recent studies have reported systems integrating load cells with microcontrollers, automated data acquisition, cloud storage, and irrigation-automation capabilities (Arruda Junior et al., 2023; Dong and Hansen, 2023; Sagar et al., 2022; McCauely and Nackley, 2022). Recent field implementations have also emphasized the importance of appropriate design and calibration for data-scarce environments and of evaluating water-balance closure and field performance beyond calibration linearity alone (Kebede et al., 2024; Lyles et al., 2024; Gupta et al., 2025). The La Libertad system addresses several of these engineering requirements through individual load-cell calibration, in situ verification using known incremental loads, an isolated drainage configuration, autonomous photovoltaic power, and real-time telemetry. However, formal analyses of measurement uncertainty, long-term drift, environmental sensitivity, and water-balance closure remain necessary during subsequent continuous field operation.

Moreover, while this study focuses on system design, calibration, and initial field verification, the high-frequency dataset generated during subsequent operation may support the development and evaluation of data-driven ET₀ models. Recent studies have combined IoT-based meteorological monitoring with hybrid ensemble machine-learning methods for reference evapotranspiration estimation (Bashir et al., 2023) and dynamic Bayesian approaches for crop evapotranspiration prediction (Kocian et al., 2023). Although model development was outside the scope of the present study, lysimeter observations may provide an independent field reference for future model training and validation in the Colombian Orinoquía.

4 Conclusion

This study addresses the critical need for direct, high-resolution measurements of crop evapotranspiration under Colombian equatorial conditions. The developed system integrates robust structural components, calibrated load cells, soil water retention analysis, and a solar-powered data transmission module to support autonomous field monitoring. Laboratory calibration produced highly linear load-cell responses, with coefficients of determination greater than 0.99, using a testing press calibrated by an ISO/IEC 17025:2017-accredited laboratory. In addition, Arachis pintoi cv. Centauro was selected as a permanent ground cover to provide a uniform and physiologically stable vegetation layer for subsequent evapotranspiration monitoring. However, crop-specific evapotranspiration performance was not evaluated in the present study.

Overall, the developed automated weighing lysimeter provides an integrated and field-deployable platform for monitoring soil–water dynamics under humid tropical conditions. Its calibrated weighing system, reconstructed soil profile, automated irrigation and drainage components, meteorological instrumentation, autonomous photovoltaic power supply, and real-time telemetry establish the technical basis for subsequent direct evapotranspiration monitoring and for evaluating model-based estimates, including FAO-56 Penman–Monteith. By enabling the generation of site-specific, high-frequency environmental observations in a data-scarce region, the system can support more robust irrigation assessments, water-resource studies, and climate-smart agricultural research in the Colombian Orinoquía and comparable tropical environments.

Future improvements to the system should focus on increasing measurement resolution and facilitating long-term maintenance. Replacing the current C3 load cells with higher-metrological-class C6 load cells of an appropriate nominal capacity could improve the detection of small mass variations associated with short-term evapotranspiration processes. In addition, future designs should incorporate a dedicated inspection duct or maintenance chamber to provide easier access to the load cells, drainage components, and electrical connections without disturbing the reconstructed soil profile. These modifications would improve both the metrological performance and long-term serviceability of the lysimeter.

References

• 1

  AllenR. G.PereiraL. S.RaesD.SmithM. (2006). Evapotranspiración del cultivo: guías para la determinación de los requerimientos de agua de los cultivos. Estudio FAO Riego Y Drenaje No. 56. Roma: FAO.

  • Google Scholar
  • View reference in article

• 2

  Arruda JuniorA.da SilvaT. J. A. (2023). Construction, calibration and performance evaluation of a lysimeter, with a double weighing system: hydraulic and mechanical. Smart Agric. Technol.4, 100222. 10.1016/j.atech.2023.100222

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  Arruda JuniorA.da SilvaT. J. A.AndradeS. P. (2023). Smart IoT lysimetry system by weighing with automatic cloud data storage. Smart Agric. Technol.4, 100177. 10.1016/j.atech.2023.100177

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BashirR. N.SaeedM.Al-SaremM.MarieR.FaheemM.KarrarA. E.et al (2023). Smart reference evapotranspiration using internet of things and hybrid ensemble machine learning approach. Internet Things24, 100962. 10.1016/j.iot.2023.100962

  • CrossRef
  • Google Scholar

• 5

  CaceresS (2025). Soil moisture, LAI, and ETo data for the Orinoquía Region. Mendeley Data. 10.17632/5m2j5kdmyt.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  Calderón CastañoL. F. (2017). Estimación y análisis de la evapotranspiración en el municipio de Manizales [master’s thesis]. Manizales: Universidad Nacional de Colombia. Available online at: https://repositorio.unal.edu.co/items/32b1fae7-090c-42d0-8d67-67ca166ff9da (Accessed March 31 2025).

  • Google Scholar
  • View reference in article

• 7

  Campbell Scientific (2026). CR1000X measurement and control datalogger. Available online at: https://www.campbellsci.ca/cr1000x (Accessed October 8 2024).

  • Google Scholar
  • View reference in article

• 8

  CarvalhoD. F.SilvaL. D. B.GuerraJ. G. M.CruzF. A.SouzaA. P. (2007). Instalação, calibração e funcionamento de um lisímetro de pesagem. Eng. Agrícola27 (2), 363–372. 10.1590/S0100-69162007000300005

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  Corporación Colombiana de Investigación Agropecuaria (AGROSAVIA) (2025). Centro de Investigación La Libertad.

  • Google Scholar
  • View reference in article

• 10

  Devia AcostaC. Y.Piñeros LizarazoR. (2021). Dinámica territorial del extractivismo agrícola y petrolero a comienzos del siglo XXI en el departamento del Meta, Colombia. Perspect. Geográfica26 (1), 37–62. 10.19053/01233769.11106

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  DongY.HansenH. (2023). Development and design of an affordable field scale weighing lysimeter using a microcontroller system. Smart Agric. Technol.4, 100147. 10.1016/j.atech.2022.100147

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  Empresas Municipales de Cali (EMCALI EICE ESP) (2026). NDC-SE-GE-003/V3.0: Geosintéticos. Norma Técnica de Servicio, Version 3.0. Cali, Colombia: EMCALI EICE ESP. Available online at: https://www.emcali.com.co/documents/d/guest/ndc-se-ge-003-v3-0 (Accessed March 15 2025).

  • Google Scholar
  • View reference in article

• 13

  Ferguson Structural Engineering Laboratory (FSEL) (2016). “Procedure for calibration of load cells,” in Rev. 0. Austin, TX: The University of Texas at Austin.

  • Google Scholar
  • View reference in article

• 14

  Gill Instruments (2022). MetConnect One: compact multi-parameter weather station—datasheet. Available online at: https://gillinstruments.com/compare-weather-stations/metconnect-weather-station/#datasheets (Accessed October 8 2024).

  • Google Scholar
  • View reference in article

• 15

  Gomes JuniorD. G.Dal PaiA.Dal PaiE. (2019). Construção de lisímetro de pesagem de seção circular para a medição da evapotranspiração de referência (ET₀): sistema com três células de carga. Irriga24 (3), 486–499. 10.15809/irriga.2019v24n3p486-499

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  Gómez BlancoJ. A.CadenaM. C. (2018). Validación de las fórmulas de evapotranspiración de referencia (ETo) para Colombia. Bogotá, DC: IDEAM. Available online at: https://www.studocu.com/co/document/universidad-de-los-andes-colombia/hidraulica/evapotranspiracion-de-referencia-eto-para-colombia/76513414?sid=7db33fd6-80b7-4594-896c-fd9d30a3b8ce1785263213 (Accessed March 31 2025).

  • Google Scholar
  • View reference in article

• 17

  GuillaumeB.Aroui BoukbidaH.BakkerG.BieganowskiA.BrostauxY.CornelisW.et al (2023). Reproducibility of the wet part of the soil water retention curve: a European interlaboratory comparison. SOIL9, 365–379. 10.5194/soil-9-365-2023

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  GuptaA.SinghR. K.KumarM. (2025). Advancing precision irrigation through an affordable IoT-enabled lysimeter for monitoring crop water requirements. Front. Sustain. Food Syst.9, 1677654. 10.3389/fsufs.2025.1677654

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  HaoX.ZhangS.LiW.DuanW.FangG.ZhangY.et al (2018). The uncertainty of Penman–Monteith method and the energy balance closure problem. J. Geophys. Res. Atmos.123, 7433–7443. 10.1029/2018JD028371

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  HenryJ. (2005). “Tropical and equatorial climates,” in Encyclopedia of World Climatology. Editor OliverJ. E. (Dordrecht: Springer), 742–750. 10.1007/1-4020-3266-8_212

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  HowellT. A.SchneiderA. D.JensenM. E. (1991). “History of lysimeter design and use for evapotranspiration measurements,” in Lysimeters for Evapotranspiration and Environmental Measurements: Proceedings of the International Symposium on Lysimetry, Honolulu, Hawaii, July 23–25, 1991. Editors AllenR. G.HowellT. A.PruittW. O.WalterI. A.JensenM. E. (American Society of Civil Engineers), 1–9. Available online at: https://www.researchgate.net/publication/279560720_History_of_Lysimeter_Design_and_Use_for_Evapotranspiration_Measurements (Accessed July 31 2025).

  • Google Scholar
  • View reference in article

• 22

  Instituto Colombiano de Normas Técnicas y Certificación (ICONTEC) (2000). NTC 174:2000. Concretos. Especificaciones de los agregados para concreto. Bogotá, D.C.: ICONTEC. Available online at: https://recordcenter.sgc.gov.co/B7/21003010024424/Documento/Nativos/Apoyo/ANEXOS/Anexo_05/NTC174.pdf. (Accessed March 15 2025).

  • Google Scholar
  • View reference in article

• 23

  Instituto de Hidrología, Meteorología y Estudios Ambientales (IDEAM) (2023). Estudio Nacional del Agua 2022. Bogotá, D.C.: IDEAM.

  • Google Scholar
  • View reference in article

• 24

  Instituto Geográfico Agustín Codazzi (IGAC) (2006). Métodos analíticos del laboratorio de suelos. 6th edBogotá, D.C.: IGAC.

  • Google Scholar
  • View reference in article

• 25

  Jaramillo RobledoA. (2006). Evapotranspiración de referencia en la región Andina de Colombia. Cenicafé57 (4), 288–298. Available online at: https://biblioteca.cenicafe.org/bitstream/10778/232/1/arc057(04)288-298.pdf. (Accessed March 31 2025)

  • Google Scholar
  • View reference in article

• 26

  KandraB.TallA.GombošM.PavelkováD. (2023). Quantification of evapotranspiration by calculations and measurements using a lysimeter. Water15 (2), 373. 10.3390/w15020373

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  KebedeN.AyanaM.MekonnenB. M.BeriH. (2024). Design, construction, and calibration of a weighable lysimeter for measuring water requirements of field crops for data-scarce areas of Ethiopia. Heliyon10, e36968. 10.1016/j.heliyon.2024.e36968

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  KocianA.CarmassiG.CelaF.ChessaS.MilazzoP.IncrocciL. (2023). IoT-based dynamic Bayesian prediction of crop evapotranspiration in soilless cultivations. Comput. Electron. Agric.205, 107608. 10.1016/j.compag.2022.107608

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  LylesB. F.SionB. D.PageD.CrewsJ. B.McDonaldE. V.HausnerM. B. (2024). Closing the water balance with a precision small-scale field lysimeter. Sensors24 (7), 2039. 10.3390/s24072039

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  McCauelyD. M.NackleyL. L. (2022). Development of mini-lysimeter system for use in irrigation automation of container-grown crops. HardwareX11, e00298. 10.1016/j.ohx.2022.e00298

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  Mendoza CastiblancoC. J.Peña QuiñonesA. J. (2021). Reference evapotranspiration estimation by different methods for the sucroenergy sector of Colombia. Rev. Bras. Eng. Agrícola Ambient.25 (9), 583–590. 10.1590/1807-1929/agriambi.v25n9p583-590

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  METER Group (2024). TEROS 12 soil moisture, temperature, and electrical conductivity sensor. Available online at: https://metergroup.com/products/teros-12/ (Accessed October 8 2024).

  • Google Scholar
  • View reference in article

• 33

  NascimentoE. F.CampecheL. F. S. M.BassoiL. H.SilvaJ. A.LimaA. C. M.PereiraF. A. C. (2011). Construção e calibração de lisímetros de pesagem para determinação da evapotranspiração e coeficiente de cultivo em videira de vinho cv. Syrah. Irriga16 (3), 271–287. 10.15809/irriga.2011v16n3p271

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  SagarA.HasanM.SinghD. K.Al-AnsariN.ChakrabortyD.SinghM. C.et al (2022). Development of smart weighing lysimeter for measuring evapotranspiration and developing crop coefficient for greenhouse chrysanthemum. Sensors22 (16), 6239. 10.3390/s22166239

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  SantiagoA. V. (2002). Evapotranspiração de referência medida por lisímetro de pesagem e estimada por Penman–Monteith (FAO-56), nas escalas mensal e decendial [master’s dissertation]. Universidade de São Paulo. Available online at: https://bdtd.ibict.br/vufind/Record/USP_8c4f08e90f52bd37d46a1f6c17d40a51 (Accessed October 8 2024).

  • Google Scholar
  • View reference in article

• 36

  Natural Resources Conservation Service (2022). Soil Survey Staff. Keys to Soil Taxonomy. 13th ed.Washington, DC: U.S. Department of Agriculture, Natural Resources Conservation Service; Available online at: https://www.nrcs.usda.gov/sites/default/files/2022-09/Keys-to-Soil-Taxonomy.pdf. (Accessed March 31 2025).

  • Google Scholar
  • View reference in article