The rapid expansion of perennial crop acreage in the past two decades raises concerns about increasing and hardening water demand . As perennial crops have a sizeable initial investment cost, fallowing perennial crops during drought results in greater economic losses than fallowing annual crops. Therefore, GSAs with a high percentage of water use by perennials will likely experience challenges in implementing sustainability management during drought. To minimize economic loss during a severe multiyear drought, a large buffer between the sustainability threshold and actual water level should be maintained for those GSAs, that is, to prepare for a 5-year drought with conditions like 2014s, a groundwater buffer should be aimed with an approximate depth at least five times the values shown in Figure 13c, after accounting for surface water availability and total porosity of the aquifer. The NASA ECOSTRESS mission has an ongoing partnership with USDA, states of California, Florida, and Iowa, and many water districts on using remote sensing evapotranspiration to better inform water resources management . Anderson et al. also find that the improved temporal frequency of ECOSTRESS resulted in improved evapotranspiration estimates and captured peak growing season during which there were no Landsat acquisitions. The mission adopted both PT-JPL and DisALEXI to map evapotranspiration ,growing berries in containers which serves as the basis for the derive L4 products, such as Water Use Efficiency and Evaporative Stress Index . ECOSTRESS also operationally provides the Priestley-Taylor potential evapotranspiration product, which has been demonstrated to be useful for water management agencies for spatial estimates of reference evapotranspiration .
Our model evaluation work suggests that PT-JPL’s evapotranspiration estimates could potentially be further improved over irrigated croplands in agricultural regions with ample evapotranspiration measures over diverse crop types. On the other hand, our spatial estimates show that EToF, which is analogous to ECOSTRESS’s Evaporative Stress Index, varies by crop types and within the Central Valley for the same crop types. Users of the Evaporative Stress Index product over Central Valley should also account for the threshold of water stress dependencies and variability by crop types and other factors such as orchard age and salinity.Although overall our refined evapotranspiration estimation approach here has similar performance to that in more complex models such as DisALEXI , there are still a few factors that can cause errors in our estimates. PT-UCD is a single-source approach. We noticed an overestimation of rice net radiation when the field was flooded, probably due to the challenges posed by heat storage in the water column and the effect of a wet surface, and an underestimation of net radiation over the two AmeriFlux corn sites between every April and July. We also found a larger uncertainty in estimating actual Priestley-Taylor coefficients for corn during the dormant season between January and April. Although evapotranspiration during the nongrowing season accounts for a small fraction of total annual water use, explicit consideration of soil, and plant components of energy balance is expected to improve the accuracy of evapotranspiration estimates. The energy balance closure issue, for example, has been well recognized for the eddy covariance measurements . Baldocchi et al. and Eichelmann et al. , for examplee.g., conducted several analyses at AmeriFlux sites in the Sacramento-San Joaquin Delta and found a closure of 79.3% in an alfalfa site, that is, the ratio of the sum of sensible and latent heat flux over the residual between net radiation and ground heat flux and storage, and 71% in a rice site.
Their study suggests that incomplete storage calculation, rather than underestimation of eddy covariance measurements of fluxes, plays a major role in the lack of energy balance closure for their sites in the Sacramento-San Joaquin Delta. Therefore, we did not perform the correction for eddy covariance measurement sites from our data set. Additional uncertainties can be introduced due to the varying footprint size of the flux towers and the scale mismatch between 30 m ET and tower measurements. We compared the annual evapotranspiration value extracted from a single Landsat pixel collocated with measurement sites and the mean of the values within the surrounding larger areas . The differences in estimates between a 90 × 90 m window and the single pixel were small, with the largest difference of −3.2% occurred at site 18 in 2016. The differences generally increased with a larger footprint, depending on the heterogeneity of the areas, e.g., over a 510 × 510 m window at site 16 in 2016 had the largest mean relative difference of −10.3% from the corresponding center pixel value. Other studies also suggested that a rigorous footprint approximation is needed in future studies to make a fair comparison with field measurements . For example, Kljun et al. ‘s flux footprint model could be implemented at flux tower sites to determine the weight and extent of the pixel window. Compared to those driven by MODIS with daily revisiting , the evapotranspiration estimates derived from Landsat have the benefit of capturing finer spatial details, which is critical for water use assessment over a heterogeneous landscape. Landsat’s 16-days repeating cycle, however, can potentially lead to uncertainty in water use monitoring, especially during the rainy season or during the rapid plant growth and senescence stages. In this study, the uncertainty due to the temporal interpolation of the missing days was found minimal overall, likely because the rainy season coincides with dormancy or the very beginning or ending of the crop growth for the majority of crops in Central Valley due to its Mediterranean climate .
There were situations, for example, right after irrigation or right after harvesting for crops that undergo multiple harvests , when a relatively large error was introduced from the estimates interpolated from observations a few days apart. Future work is also needed to increase the temporal resolution of water use estimate by fuzing Sentinel 2 A&B satellite observations every 5 days with Landsat data. A sophisticated data fusion technique will also improve the accuracy of evapotranspiration monitoring and assessment, by taking advantage of complementary observations from multiple sensors . The robustness of our optimization approach partly relies on the availability of multiple field measurements for diverse crop types across the Central Valley. The automatic workflow developed here allows for a continuous improvement of the optimization accuracy, by taking advantages of the increasingly available crop evapotranspiration measurements with the increased deployment of both research-grade and commercialized surface renewal stations in the state . Although the Priestley-Taylor parameters in this study were tailored for California’s cropland, our data-model integration framework is generalizable to other regions. Once recalibrated and tested with local field data,blueberry container the PT-UCD approach can be applied to monitor daily evapotranspiration and assess water use at various scales over regions besides the Central Valley.Organophosphate pesticides are commonly used insecticides that inhibit acetylcholinesterase enzyme function and have been associated with poorer neurodevelopment in children . Children are particularly susceptible to the adverse impacts of pesticides and those living in agricultural areas may be exposed via multiple pathways, including diet, drinking water, residential use, drift from agricultural applications, and take-home exposures . Assessing exposure to OP pesticides is difficult due to their short biologic half lives and rapid excretion from the body . Dialkylphosphate metabolites, the most commonly used biomarker to characterize OP exposure in epidemiologic studies , have biological half-lives of less than 30 min to >24 h, depending on the parent OP and route of exposure . Measurements of metabolites or parent chemicals in 24-hr urine samples are considered the “gold standard” for assessing daily exposure to pesticides and other environmental chemicals that are excreted in urine . However, factors such as cost and participant burden make it difficult to collect 24-hr samples . While collection of spot urine samples is a convenient alternative, research suggests that analysis of biomarkers with short half-lives, including DAPs, in spot samples may result in exposure misclassification due to higher inter- and intra-individual variability . First morning void urine samples may reduce exposure misclassification, as they are more concentrated and reflect a longer period of accumulation . Few studies have assessed how well either random spot or FMV urine samples approximate internal pesticide dose estimated from 24-hr samples, information that is critical for risk assessment and pesticide regulation.
Estimating dose based on metabolite concentrations from spot samples also requires an accurate measure of urinary dilution and total daily urinary output volume . In adults, 24-hr urinary metabolite excretion has been estimated from spot urine samples by adjusting for creatinine excretion as an index of total daily urinary output volume . However, few studies have evaluated the validity of this approach in children. Due to likely differences in children’s urinary creatinine excretion from factors including age, sex, muscle mass, body mass index , diet, and fluid intake , adjusting for creatinine to estimate toxicant doses in children may introduce unknown sources of variability . Although not used as widely as creatinine correction, some evidence suggests that adjusting for specific gravity may be a more robust method to account for urinary output among children . The US Environmental Protection Agency is mandated by the 1996 Food Quality Protection Act to review and establish health-based standards for pesticide residues in foods and examine the cumulative health effects of exposure to mixtures of pesticides that share a common mechanism of toxicity, with prioritization of pesticides that may pose the greatest risk, such as OPs . The U.S. EPA has selected the Relative Potency Factor method to conduct hazard and dose-response assessments. RPFs are calculated as the ratio of the toxic potency of a given chemical, determined by the oral benchmark dose10 value based on a 10% brain cholinesterase inhibition, to that of an index chemical. Individual OP doses derived from index chemical toxicity equivalent doses can be summed to create cumulative OP dose equivalents . In this study, we measured DAP metabolites in spot and 24-hr void urine samples collected from 25 preschool-aged children over 7 consecutive days. The objective of this analysis was to evaluate the validity of using volume- and creatinine-adjusted FMV and non-FMV spot urine samples to estimate total 24-hr OP dose in children according to the 2006 US EPA Organophosphorus Cumulative Risk Assessment guidelines. The results of these analyses have implications for policy and risk assessments and could serve as a case study for other non-persistent toxicants measured in urine. Subject recruitment and procedures have been described previously . Briefly, we enrolled a convenience sample of 25 children recruited from clinics serving low-income families in the Salinas Valley, California. Eligible children were 3–6 years old, in good health with no history of diabetes or renal disease, toilet trained, and free of enuresis, and had English- or Spanish-speaking mothers who were ≥ 18 years old. Sampling occurred in March and April 2004. The University of California at Berkeley Committee for the Protection of Human Subjects approved all study procedures and parents provided written informed consent. Each family participated in the study over 7 consecutive days. On the first day, study staff measured the participating child’s height and weight, provided the supplies needed to collect urine samples, including specimen trays and jars, gloves, collection jars with blank labels, a small refrigerator, and two 24-hr sampling record forms, and instructed the parents and child on how to collect, record, and store samples. Urine voids were collected either directly into a collection jar or into a sterile pre-cleaned specimen tray placed over the toilet, which was then transferred by parents into the collection jar. Fig. 1 shows the timing of study activities. On spot-sampling days , families collected a single void at their convenience, recording the time of collection on the jar labels and identifying the sample as an FMV or non-FMV spot sample. On 24-hr sampling days , families were instructed to collect all urine voids from the 24-hr period as separate specimens, including the child’s FMV, all daytime and evening spot voids, and the FMV of the following day , if it occurred within the 24-hr sampling period. Participants were instructed to record the timing of all voids, including missed voids, on the 24-hr sampling record form. We limited the current analyses to samples collected on 24-hr sampling days .