Reference evapotranspiration of Brazil modeled with machine learning techniques and remote sensing.

Reference evapotranspiration (ETo) is a fundamental parameter for hydrological studies and irrigation management. The Penman-Monteith method is the standard to estimate ETo and requires several meteorological elements. In developing countries, the number of weather stations is insufficient. Thus, free products of remote sensing with evapotranspiration information must be used for this purpose. In this context, the objective of this study was to estimate monthly ETo from potential evapotranspiration (PET) made available by MOD16 product. In this study, the monthly ETo estimated by Penman-Monteith method was considered as the standard. For this, data from 265 weather station of the National Institute of Meteorology (INMET), spread all over the Brazilian territory, were acquired for the period from 2000 to 2014 (15 years). For these months, monthly PET values from MOD16 product for all Brazil were also downloaded. By using machine learning algorithms and information from WorldClim as covariates, ETo was estimated through images from the MOD16 product. To perform the modeling of ETo, eight regression algorithms were tested: multiple linear regression; random forest; cubist; partial least squares; principal components regression; adaptive forward-backward greedy; generalized boosted regression and generalized linear model by likelihood-based boosting. Data from 2000 to 2012 (13 years) were used for training and data of 2013 and 2014 (2 years) were used to test the models. The PET made available by the MOD16 product showed higher values than those of ETo for different periods and climatic regions of Brazil. However, the MOD16 product showed good correlation with ETo, indicating that it can be used in ETo estimation. All models of machine learning were effective in improving the performance of the metrics evaluated. Cubist was the model that presented the best metrics for r2 (0.91), NSE (0.90) and nRMSE (8.54%) and should be preferred for ETo prediction. MOD16 product is recommended to be used to predict monthly ETo, which opens possibilities for its use in several other studies.

The first IAEA inter-laboratory comparison exercise in Latin America and the Caribbean for stable isotope analyses of water samples.

The use of stable isotopes (δ 2H and δ 18O) is widespread in water resources studies. In the Latin America and the Caribbean (LAC) region, the application of isotope techniques has increased in the past decade, but there remains room to gain self-reliance in environmental isotope studies, necessitating easy and fast access to good-quality isotope data. To that end, in 2018 the IAEA carried out the first regional interlaboratory comparison exercise, testing the analytical performance of 25 laboratories using isotope-ratio mass spectrometry and laser absorption spectroscopy. The three test samples covered a commonly observed range of 0 to -16 ‰ δ 18O and 0 to -115 ‰ δ 2H. z- and ζ-scores were used to benchmark laboratories' performance against a strict criterion. We found that 81% of the laboratories had satisfactory performance ( | z | ¯ ≤ 2) for δ 2H but only 54% achieved similar scores for δ 18O. Only a minor fraction of results (12% for δ 2H and 15% for δ 18O) were unsatisfactory. The larger number of questionable results for δ 18O confirmed the challenges in laser absorption spectroscopy for this isotope. Besides instrumental performance, the sample throughput, laboratory reference materials, and data post-processing were contributing factors to inaccurate or imprecise performance.

HSPF-based watershed-scale water quality modeling and uncertainty analysis.

This paper presents findings on uncertainties, introduced through digital elevation model (DEM) resolution and DEM resampling, in watershed-scale flow and water quality (NO3, P, and total suspended sediment) simulations. The simulations were performed using the Better Assessment Science Integrating Point and Nonpoint Sources/Hydrological Simulation Program Fortran watershed modeling system for two representative study watersheds delineated with both the original DEMs of four different resolutions (including 3.5, 10, 30, and 100 m) and the resampled DEMs of three different resolutions (including 10, 30, and 100 m), creating 14 simulation scenarios. Parameter uncertainties were quantified by means of the GLUE approach and compared to input data uncertainties. Results from the 14 simulation scenarios showed that there was a common increasing trend in errors of simulated flow and water quality parameters when the DEM resolution became coarser. The errors involved in the watershed with a mild slope were found to be substantially (up to 10 times) greater than those of the other watershed with a relatively steep slope. It was also found that sediment was the most sensitive and NO3 was the least sensitive parameters to the variation in DEM resolution, as evidenced by the maximum normalized root mean square error (NRMSE) of 250% in the simulated sediment concentration and 11% in the simulated NO3 concentration, respectively. Moreover, results achieved from the resampled (particularly coarser) DEMs were significantly different from corresponding ones from original DEMs. By comparing uncertainties from different sources, it was found that the parameter-induced uncertainties were higher than the resolution-induced uncertainties particularly in simulated NO3 and P concentrations for studied watersheds. The findings provide new insights into the sensitivity and uncertainty of water quality parameters and their simulation results, serving as the guidelines for developing and implementing water quality management and watershed restoration plans.

Evaluating the ability of standardised leaching tests to predict metal(loid) leaching from intact soil columns using size-based elemental fractionation.

Laboratory-based leaching tests are frequently used for in situ risk assessments of contaminant leaching to groundwater and surface waters. This study evaluated the ability of three standardised leaching tests to assess leaching of lead (Pb), zinc (Zn), arsenic (As) and antimony (Sb) from four intact soil profiles, by considering particulate (0.45-8 µm; percolation test), colloidal (10 kDa-0.45 µm) and truly dissolved (<10 kDa) fractions of these elements. Deionised water was used as the percolation test leachant, while either deionised water or 1 mM CaCl2 was used in batch tests. Data from an irrigation experiment were used as reference. The results indicated that in percolation tests, leachate should be collected at a liquid:solid ratio (L/S) range of 2-10, instead of 0-0.5 or 0.5-2. Even at L/S = 2-10, the percolation test overestimated total Pb concentration, mainly because of greater mobilisation of particle-bound Pb, but appeared suitable for categorising soils into high/low risk with respect to mobilisation of particulate and colloidal contaminants. The batch test performed better with CaCl2 than with deionised water when standard membrane filtration (0.45 µm) was used, as the high Ca2+ concentration reduced colloidal mobilisation, avoiding overestimation of concentrations of elements such as Pb. However, the higher Ca2+ concentration and lower pH could result in overestimated concentrations of weakly sorbed elements, e.g. Zn.

Caracterización y modelado hidrometeorológico en la cuenca del río Atulapa, Guatemala / Characterization and hydrometeorological modeling in the Atulapa River basin, Guatemala

Los modelos hidrometeorológicos, facilitan el control, monitoreo y planificación del recurso agua en cuencas hidrográficas, representando variables meteorológicas de forma distribuida. El objetivo del estudio, es proponer modelos hidrometeorológicos bidimensionales para cuantificar de forma precisa, las variables básicas que determinan la dinámica del recurso hídrico en la cuenca del río Atulapa, Esquipulas, departamento de Chiquimula, Guatemala. Se generaron y recopilaron datos hidrometeorológicos, con estaciones digitales instaladas y estaciones locales. Para el desarrollo de los modelos se elaboró una correlación entre altitud de las estaciones meteorológicas, como variable explicativa y registros climáticos como variable de respuesta. El modelo de escorrentía es una correlación entre alturas limnimétricas en la estación hidrométrica, como variable explicativa y el caudal óptimo de los aforos como variable de respuesta. La precipitación media de la cuenca obtenida con los modelos generados, es de 1,884 mm/año y la temperatura media de 18.92°C. El cambio de temperatura entre la parte baja y alta disminuye a razón de 1°C por cada 175 m.snm. El caudal medio anual es a razón de 0.5 m3 /seg en la estación hidrométrica del Puente. Los modelos hidrometeorológico generados en la cuenca del río Atulapa representaron la variabilidad climática en forma bidimensional, permitiendo un control y monitoreo del recurso hídrico en la cuenca, para la planificación del uso sostenible. La orografía en la cuenca, definió la estrecha relación con las variables meteorológicas precipitación y temperatura obteniendo modelos de regresión lineal precisos.

The hydrometeorological models facilitate the control and monitoring of the water resource in watersheds, representing meteorological variables in a distributed way, for the planning of the water resource. The objective of the study is to propose two-dimensional hydrometeorological models to accurately quantify the basic variables that determine the dynamics of water resources in the Atulapa River basin, Esquipulas, department of Chiquimula, Guatemala. Hydrometeorological data were generated and collected, with installed digital stations and local stations. For the development of the models, a correlation was made between the altitude of the meteorological stations, as an explanatory variable and climatic records as a response variable. The runoff model is a correlation between limnimetric heights in the hydrometric station, as an explanatory variable and the optimum flow of the gauging as a response variable. The average precipitation of the basin obtained with the generated models is 1,884 mm/ year and the average temperature is 18.92°C. The temperature change between the low and high part decreases at a rate of 1°C for every 175 masl. The average annual flow is at a rate of 0.5 m3 /sec at the Puente hydrometric station. The hydrometeorological models generated in the Atulapa River basin represented the climatic variability in two-dimensional form, allowing a control and monitoring of the water resource in the basin, for the planning of the sustainable use. The orography in the basin, defined the close relationship with meteorological variables precipitation and temperature obtaining accurate linear regression models.