Influence of data preprocessing on the quantitative determination of nutrient content in poultry manure by near infrared spectroscopy

With increasing concern over potential polltion from farm wastes, there is a need for rapid and robust methods that can analyze livestock manure nutrient content. The near infrared spectroscopy (NIRS) method was used to determine nutrient content in diverse poultry manure samples (n=91). Various standard preprocessing methods (derivatives, multiplicative scatter correction, Savitsky-Golay smoothing, and standard normal variate) were applied to reduce data systemic noise. In addition, a new preprocessing method known as direct orthogonal signal correction (DOSC) was tested. Calibration models for ammonium nitrogen, total potassium, total nitrogen, and total phosphorus were developed with the partial least squares (PLS) method. The results showed that all the preprocessed data improved prediction results compared with the non-preprocessing method. Compared with the other preprocessing methods, the DOSC method gave the best results. The DOSC method achieved moderately successful prediction for ammonium nitrogen, total nitrogen, and total phosphorus. However, all preprocessing methods did not provide reliable prediction for total potassium. This indicates the DOSC method, especially combined with other preprocessing methods, needs further study to allow a more complete predictive analysis of manure nutrient content.     

Soil testing to predict phosphorus leaching

Subsurface pathways can play an important role in agricultural phosphorus (P) losses that can decrease surface water quality. This study evaluated agronomic and environmental soil tests for predicting P losses in water leaching from undisturbed soils. Intact soil columns were collected for five soil types that a wide range in soil test P. The columns were leached with deionized water, the leachate analyzed for dissolved reactive phosphorus (DRP), and the soils analyzed for water-soluble phosphorus (WSP), 0.01 M CaCl2 P (CaCl2-P), iron-strip phosphorus (FeO-P), and Mehlich-1 and Mehlich-3 extractable P, Al, and Fe. The Mehlich-3 P saturation ratio (M3-PSR) was calculated as the molar ratio of Mehlich-3 extractable P/[Al + Fe]. Leachate DRP was frequently above concentrations associated with eutrophication. For the relationship between DRP in leachate and all of the soil tests used, a change point was determined, below which leachate DRP increased slowly per unit increase in soil test P, and above which leachate DRP increased rapidly. Environmental soil tests (WSP, CaCl2-P, and FeO-P) were slightly better at predicting leachate DRP than agronomic soil tests (Mehlich-1 P, Mehlich-3 P, and the M3-PSR), although the M3-PSR was as good as the environmental soil tests if two outliers were omitted. Our results support the development of Mehlich-3 P and M3-PSR categories for profitable agriculture and environmental protection; however, to most accurately characterize the risk of P loss from soil to water by leaching, soil P testing must be fully integrated with other site properties and P management practices.     

CHLOROPHYLL-BIOMASS-NUTRIENT RELATIONSHIPS FOR NATURAL ASSEMBLAGES OF FLORIDA PHYTOPLANKTON1

Parallel determination of phytoplankton biomass and chlorophyll a concentration were made on spring and summer phytoplankton samples collected from 165 Florida lakes. There was a significant correlation between chlorophyll a concentration and phytoplankton biomass (r=0.80; P < 0.01). Chlorophyll content per unit phytoplankton biomass ranged over two orders of magnitude. Nitrogen seemed to be a major factor influencing the chlorophyll content of Florida algae. Multiple regression analyses indicated that phytoplankton biomass was dependent on both the total phosphorus and total nitrogen concentration. Nutrient-phytoplankton and Secchi-phytoplankton relationships for the Florida lakes had higher coefficients of determination if chlorophyll a concentrations rather than phytoplankton biomass data were used in regression analyses.     

COMPARING SAMPLING SCHEMES FOR MONITORING POLLUTANT EXPORT FROM A DAIRY PASTURE1

ABSTRACT: Dairy cow pastures and feeding areas around barns can be a significant source of nonpoint source pollutants to nearby streams. To help document the significance of these sources, nutrient export in streamfiow from a 56.7-ha, mostly agricultural, watershed located in southwestern North Carolina was monitored from August 1994 to January 1996. Total nitrogen and phosphorus export rates from the upper, predominantly pasture, part of the watershed were 18.0 and 1.4 kg/ha/yr, respectively, as measured by weekly grab sampling and 18.7 and 4.9 kg/halyr, respectively, as measured from storm event monitoring. Nitrogen and phosphorus export rates for the area between the monitoring sites, which included overgrazed cow holding and feeding areas and farm buildings, were 376 and 86 kgfhalyr, respectively, for grab sampling and 351 and 160 kg/ha/yr, respectively, for storm event monitoring. To estimate the amount of reduction from nonpoint source controls necessary to effect a significant reduction in pollutant loading, statistical analyses of the load data were conducted. The analyses for the five pollutants monitored showed that total suspended solids would require the greatest reduction (34.6 percent for weekly grab and 33.6 percent for storm) in loading after the implementation of controls for statistical significance. Nitrate plus nitrite was found to require the least reduction (12.6 percent for weekly grab). Pollutant export rates computed from weekly grab samples and storm event samples used separately were compared to corresponding export rates computed from combining grab and storm event samples to assess the differences in monitoring schemes.     

Relationship between soil test phosphorus and phosphorus in runoff: effects of soil series variability

Phosphorus loss in runoff from agricultural fields has been identified as an important contributor to eutrophication. The objective of this research was to determine the relationship between phosphorus (P) in runoff from a benchmark soil (Cecil sandy loam; fine, kaolinitic, thermic Typic Kanhapludult) and Mehlich III-, deionized water-, and Fe(2)O(3)-extractable soil P, and degree of phosphorus saturation (DPS). Additionally, the value of including other soil properties in P loss prediction equations was evaluated. Simulated rainfall was applied (75 mm h(-1)) to 54 1-m(2) plots installed on six fields with different soil test phosphorus (STP) levels. Runoff was collected in its entirety for 30 min and analyzed for total P and dissolved reactive phosphorus (DRP). Soil samples were collected from 0- to 2-, 0- to 5-, and 0- to 10-cm depths. The strongest correlation for total P and DRP occurred with DPS (r(2) = 0.72). Normalizing DRP by runoff depth resulted in improved correlation with deionized water-extractable P for the 0- to 10-cm sampling depth (r(2) = 0.81). The STP levels were not different among sampling depths and analysis of the regression equations revealed that soil sampling depth had no effect on the relationship between STP and P in runoff. For all forms of P in runoff and STP measures, the relationship between STP and runoff P was much stronger when the data were split into groups based on the ratio of oxalate-extractable Fe to Al. For all forms of P in runoff and all STP methods, R(2) increased with the inclusion of oxalate-extractable Al and Fe in the regression equation. The results of this study indicate that inclusion of site-specific information about soil Al and Fe content can improve the relationship between STP and runoff P.     

Applying value of information methods to prioritize elements for water quality management with an example of linear alkylbenzene sulfonate in the Yodo River,Japan

Narrowing the decision space is crucial in water quality management at the meso-scale for developing countries, where a lack of data and financial budgets prevent the development of appropriate management plans and result in serious water quality degradation in many rivers. In this study, a framework for handling this task is proposed, comprising a lumped water quality model, with sensitivity and uncertainty analyses, and a management domain, including loss estimation and value of information analysis. Through a case study with linear alkylbenzene sulfonate (LAS) in the Yodo River, it is found that non-point sources and flow rate are factors that influence LAS concentration at the hot spot location. By considering the entire process of water quality management planning, we identify that the definition of the cost function of LAS treatment determines the appropriate estimation for the expected loss in reducing LAS under uncertain water quality conditions. The value of information analysis with “expected value of including uncertainty” and “expected value of perfect information” further helps estimate the benefit of including uncertainty in decision-making and the financial cost for obtaining more information regarding inputs that have been previously prioritized.     

GENERALIZED WATERSHED LOADING FUNCTIONS FOR STREAM FLOW NUTRIENTS1

ABSTRACT: Loading functions are proposed as a general model for estimating monthly nitrogen and phosphorus fluxes in stream flow. The functions have a simple mathematical structure, describe a wide range of rural and urban nonpoint sources, and couple surface runoff and ground water discharge. Rural runoff loads are computed from daily runoff and erosion and monthly sediment yield calculations. Urban runoff loads are based on daily nutrient accumulation rates and exponential wash off functions. Ground water discharge is determined by lumped parameter unsaturated and saturated zone soil moisture balances. Default values for model chemical parameters were estimated from literature values. Validation studies over a three-year period for an 850 km² watershed showed that the loading functions explained at least 90 percent of the observed monthly variation in dissolved and total nitrogen and phosphorus fluxes in stream flow. Errors in model predictions of mean monthly fluxes were: dissolved phosphorus - 4 percent; total phosphorus - 2 percent; dissolved nitrogen - 18 percent; and total nitrogen - 28 percent. These results were obtained without model calibration.     

ANTIDEGRADATION WATER QUALITY CRITERIA FOR THE DELAWARE RIVER: A DISTRIBUTION-FREE STATISTICAL APPROACH1

ABSTRACT: Existing water quality for the Middle Delaware Scenic and Recreational River is significantly better than is required by current standards, leaving a potential for degradation. A method is presented for deriving candidate antidegradation water quality criteria for this segment of the Delaware River using statistical analysis of historic (ambient) water quality data. Data for 34 water quality parameters are first evaluated for data density, serial correlation, trend, seasonality, and other factors. These preliminary analyses are based on observation of data plots and application of distribution-free statistical techniques that are insensitive to outliers and are robust to relatively mild violations of basic assumptions. Data for 12 of the parameters have sufficient density for further analysis and can reasonably be modeled as independent and identically distributed over time (either seasonally or for the entire data sets). For these 12 parameters, distribution-free statistical methods are developed and used to derive intervals within which there is high confidence (usually greater than 95 percent) that the quantiles with potential use as anti-degradation criteria (the 0.85^th, 0.90^th, and 0.95^th quantiles in this study) for a particular parameter lie.     

Modeling the relationship between land use and surface water quality

It is widely known that watershed hydrology is dependent on many factors, including land use, climate, and soil conditions. But the relative impacts of different types of land use on the surface water are yet to be ascertained and quantified. This research attempted to use a comprehensive approach to examine the hydrologic effects of land use at both a regional and a local scale. Statistical and spatial analyses were employed to examine the statistical and spatial relationships of land use and the flow and water quality in receiving waters on a regional scale in the State of Ohio. Besides, a widely accepted watershed-based water quality assessment tool, the Better Assessment Science Integrating Point and Nonpoint Sources (BASINS), was adopted to model the plausible effects of land use on water quality in a local watershed in the East Fork Little Miami River Basin. The results from the statistical analyses revealed that there was a significant relationship between land use and in-stream water quality, especially for nitrogen, phosphorus and Fecal coliform. The geographic information systems (GIS) spatial analyses identified the watersheds that have high levels of contaminants and percentages of agricultural and urban lands. Furthermore, the hydrologic and water quality modeling showed that agricultural and impervious urban lands produced a much higher level of nitrogen and phosphorus than other land surfaces. From this research, it seems that the approach adopted in this study is comprehensive, covering both the regional and local scales. It also reveals that BASINS is a very useful and reliable tool, capable of characterizing the flow and water quality conditions for the study area under different watershed scales. With little modification, these models should be able to adapt to other watersheds or to simulate other contaminants. They also can be used to study the plausible impacts of global environmental change. In addition, the information on the hydrologic effects of land use is very useful. It can provide guidelines not only for resource managers in restoring our aquatic ecosystems, but also for local planners in devising viable and ecologically-sound watershed development plans, as well as for policy makers in evaluating alternate land management decisions.     

ANTIDEGRADATION WATER QUALITY CRITERIA FOR THE DELAWARE RIVER: A DISTRIBUTION-FREE STATISTICAL APPROACH1

ABSTRACT: Existing water quality for the Middle Delaware Scenic and Recreational River is significantly better than is required by current standards, leaving a potential for degradation. A method is presented for deriving candidate antidegradation water quality criteria for this segment of the Delaware River using statistical analysis of historic (ambient) water quality data. Data for 34 water quality parameters are first evaluated for data density, serial correlation, trend, seasonality, and other factors. These preliminary analyses are based on observation of data plots and application of distribution-free statistical techniques that are insensitive to outliers and are robust to relatively mild violations of basic assumptions. Data for 12 of the parameters have sufficient density for further analysis and can reasonably be modeled as independent and identically distributed over time (either seasonally or for the entire data sets). For these 12 parameters, distribution-free statistical methods are developed and used to derive intervals within which there is high confidence (usually greater than 95 percent) that the quantiles with potential use as anti-degradation criteria (the 0.85^th, 0.90^th, and 0.95^th quantiles in this study) for a particular parameter lie.     

DISTRIBUTION OF NITRATE AND ORTHOPHOSPHATE IN SELECTED STREAMS IN CENTRAL NEBRASKA1

ABSTRACT: The Central Nebraska Basins is one of 60 study units in the National Water-Quality Assessment Program of the U.S. Geological Survey. The study unit includes the Platte River and two major tributaries, the Loup and Elkhorn Rivers. Agriculture is the predominant land use in the study unit, with only eight urbanized communities exceeding a population of 10,000. Water samples were collected from selected streams in the study unit during 1993–1995. The data were used to assess the distribution of nitrogen compounds and phosphorus in the streams and to relate the concentrations of these constituents to environmental settings. This article focuses on dissolved nitrate and orthophosphate. Dissolved nitrate concentrations were highest (90th percentiles were less than 7.0 milligrams per liter as nitrogen) in areas with extensive cropland and pasture, where chemical fertilizers are intensively applied. Synoptic measurements conducted in March and August 1994 indicate that relatively little residual fertilizer, as nitrate, applied during a single crop-growing season enters streams. Dissolved nitrate concentrations showed a seasonal pattern, being highest during winter months and lowest during the late spring and summer. Dissolved orthophosphate concentrations tended to be low across the study unit, 90 percent of all analyses did not exceed 1.7 milligrams per liter as phosphorus.     

Speciation of phosphorus in phosphorus-enriched agricultural soils using X-ray absorption near-edge structure spectroscopy and chemical fractionation

Knowledge of phosphorus (P) species in P-rich soils is useful for assessing P mobility and potential transfer to ground water and surface waters. Soil P was studied using synchrotron X-ray absorption near-edge structure (XANES) spectroscopy (a nondestructive chemical-speciation technique) and sequential chemical fractionation. The objective was to determine the chemical speciation of P in long-term-fertilized, P-rich soils differing in pH, clay, and organic matter contents. Samples of three slightly acidic (pH 5.5-6.2) and two slightly alkaline (pH 7.4-7.6) soils were collected from A or B horizons in two distinct agrosystems in the province of Québec, Canada. The soils contained between 800 and 2100 mg total P kg(-1). Distinct XANES features for Ca-phosphate mineral standards and for standards of adsorbed phosphate made it possible to differentiate these forms of P in the soil samples. The XANES results indicated that phosphate adsorbed on Fe- or Al-oxide minerals was present in all soils, with a higher proportion in acidic than in slightly alkaline samples. Calcium phosphate also occurred in all soils, regardless of pH. In agreement with chemical fractionation results, XANES data showed that Ca-phosphates were the dominant P forms in one acidic (pH 5.5) and in the two slightly alkaline (pH 7.4-7.6) soil samples. X-ray absorption near-edge structure spectroscopy directly identified certain forms of soil P, while chemical fractionation provided indirect supporting data and gave insights on additional forms of P such as organic pools that were not accounted for by the XANES analyses.     

Using rank-order geostatistics for spatial interpolation of highly skewed data in a heavy-metal contaminated site

The spatial distribution of a pollutant in contaminated soils is usually highly skewed. As a result, the sample variogram often differs considerably from its regional counterpart and the geostatistical interpolation is hindered. In this study, rank-order geostatistics with standardized rank transformation was used for the spatial interpolation of pollutants with a highly skewed distribution in contaminated soils when commonly used nonlinear methods, such as logarithmic and normal-scored transformations, are not suitable. A real data set of soil Cd concentrations with great variation and high skewness in a contaminated site of Taiwan was used for illustration. The spatial dependence of ranks transformed from Cd concentrations was identified and kriging estimation was readily performed in the standardized-rank space. The estimated standardized rank was back-transformed into the concentration space using the middle point model within a standardized-rank interval of the empirical distribution function (EDF). The spatial distribution of Cd concentrations was then obtained. The probability of Cd concentration being higher than a given cutoff value also can be estimated by using the estimated distribution of standardized ranks. The contour maps of Cd concentrations and the probabilities of Cd concentrations being higher than the cutoff value can be simultaneously used for delineation of hazardous areas of contaminated soils.     

Phosphorus in runoff assessed by anion exchange resin extraction and an algal assay

Eutrophication of surface waters can be accelerated by agricultural inputs of phosphorus (P), provided that P is in a form that can be utilized by aquatic algae. We studied anion exchange resin (AER) extraction and a dual culture algal assay (DCAA) for the determination of potentially algal-available P in water samples without sediment preconcentration. Our material consisted of agricultural and forest runoff and wastewaters. The results obtained by the two methods were essentially equal when the samples contained only small amounts of particulate phosphorus (PP) in relation to dissolved molybdate-reactive phosphorus (DRP). However, in turbid agricultural runoff, P extracted with AER averaged 72% (n = 17) of the P yield of the 3-wk DCAA (R2 = 0.94). When the runoff samples were diluted for the AER extraction in the same manner as for the DCAA, the AER-P yield increased to 85% (n = 5) of DCAA-P. The minimum detectable value was greater for the AER test (41 microg L(-1) AER-extractable P) than for the DCAA (7 microg L(-1) DCAA-P). At concentrations greater than about 50 microg L(-1) AER-P or DCAA-P, the accuracy of the methods was satisfactory, with the coefficient of variation in replicated analyses being less than 10% for the AER test and less than 20% for the DCAA. Other anions competing for the exchange sites of the AER decreased P recovery by 15 to 20% when their equivalent concentration exceeded about 4 mmol, L(-1), and this effect was relatively constant over a large concentration range. We consider that AER extraction is a suitable low-cost method to estimate the algal availability of P in runoff samples.     

TROPICAL EUTROPHIC LAKE MODELING AND PARAMETER DISCRIMINATION1

ABSTRACT: Water quality data collected between 1978 and 1981 in a highly lake in Northern Venezuela, Lake Valencia, were analyzed to detect spatial and temporal trends. Based on the results of the analyses, an appropriate nutrient-algae dynamics model was formulated. Because many parameters, such as the algae concentration were constant over time, and the model is time dependent, the model had to be calibrated with the use of a large and structured trial-and-error calibration process. Through the calibration process, the most sensitive parameters of the model were identified, and are in order of importance: the chlorophyll-to-nitrogen ratio for algae, the algae settling velocity, the phosphorus release rate from the sediments, the chlorophyll-to-phosphorus ratio for algae, and the exchange coefficient in the upper layer of the lake. Model simulations showed that a reduction in the nitrogen load to the lake as well as a reduction in the phosphorus load will decrease the algae population. These model simulations had a high degree of uncertainty associated with them, making additional sampling directed towards the measurement of the sensitive parameters desirable.     

A FRAMEWORK FOR PHOSPHORUS TRANSPORT MODELING IN THE LAKE OKEECHOBEE WATERSHED1

ABSTRACT: A modeling framework was developed to determine phosphorus loadings to Lake Okeechobee from watersheds located north of the lake. This framework consists of the land-based model CREAMS-WT, the in-stream transport model QUAL2E, and an interface procedure to format the land-based model output for use by the in-stream model. QUAL2E hydraulics and water quality routines were modified to account for flow routing and phosphorus retention in both wetlands and stream channels. Phosphorus loadings obtained from previous applications of CREAMS-WT were used by QUAL2E, and calibration and verification showed that QUAL2E accurately simulated seasonal and annual phosphorus loadings from a watershed. Sensitivity and uncertainty analyses indicated that the accuracy of monthly loadings can be improved by using better estimates of in-stream phosphorus decay rates, ground water phosphorus concentrations, and runoff phosphorus concentrations as input to QUAL2E.     

VERIFICATION AND APPLICATION OF REGIONAL EQUATIONS FOR THE STORAGE FUNCTION RUNOFF MODEL1

The storage function model is a nonlinear rainfall-runoff model that has been developed for and applied to flood runoff analysis in Japan. This paper extends the model applicability by developing practical equations for estimating model parameters which are appropriate on a regional basis, i.e., so-called regional equations. Previously, the parameters were computed from historical data for a specific basin or from relationships that do not account for land use and topography. To develop the regionalized equations, model parameters were identified for 91 flood events from 22 watersheds in Japan by applying a mathematical optimization technique. Results from 39 of these events were statistically compared and regional relationships were determined as a function of land use, basin area and rainfall intensity. The utility of the estimated equations were tested by computing runoff hydrographs for lumped basins. The estimated parameters were also applied in a distributed watershed model formulation. Both applications showed acceptable results that validate the use of the regionalized relationships.     

Determination of redox-sensitive phosphorus in field runoff without sediment preconcentration

Reduction-induced phosphorus (P) release from particles transported by field runoff has been poorly studied for want of a method that could be used for large surveys. To rectify this shortcoming, we modified the bicarbonate-dithionite (BD) extraction step of a sediment P speciation scheme for analyzing redox-sensitive P in runoff without sample preconcentration. The extraction comprised the addition of bicarbonate (pH buffer) and dithionite (reducing agent) into a runoff sample, 15 min of gentle shaking, filtration, and sample digestion. The samples were greatly reduced (Eh < -200 mV), and Fe and P were solubilized, but Al solubility was not increased. Phosphorus release from rock phosphates (calcium phosphates) was greater in the BD extraction than in water or bicarbonate solution, although no more than 0.2% of the total P was released. For runoff from a very fine Typic Cryaquept, the particulate phosphorus (PP) versus BD-PP relationship was linear up to a PP concentration of about 1.0 mg L(-1), but over the whole PP range studied (up to 2.6 mg L(-1)) somewhat better described by an exponential equation (BD-PP = 0.297 x PP(0.766); r2 = 0.91, n = 79). The minimum detectable value given by the method was relatively low, 0.023 mg L(-1), but reproducibility varied, with the coefficient of variation for 10 samples analyzed with 5 replicates ranging from 1.8 to 28.5%. Considering the variable reproducibility of the results and the lack of suitable reference material, the method needs further refinement and testing if it is to be used for quantitative determination of redox-sensitive P in runoff.     

Agricultural Land Use and Best Management Practices to Control Nonpoint Water Pollution

In recent years, improvements in point-source depuration technologies have highlighted the problems regarding agricultural nonpoint (diffuse) sources, and this issue has become highly relevant from the environmental point of view. The considerable extension of the areas responsible for this kind of pollution, together with the scarcity of funds available to local managers, make minimizing the impacts of nonpoint sources on a whole basin a virtually impossible task. This article presents the results of a study intended to pinpoint those agricultural areas, within a basin, that contribute most to water pollution, so that operations aimed at preventing and/or reducing this kind of pollution can be focused on them. With this aim, an innovative approach is presented that integrates a field-scale management model, a simple regression model, and a geographic information system (GIS). The Lake Vico basin, where recent studies highlighted a considerable increase in the trophic state, mainly caused by phosphorus (P) compounds deriving principally from the intensive cultivation of hazelnut trees in the lake basin, was chosen as the study site. Using the management model Groundwater Loading Effects of Agricultural Management Systems (GLEAMS), the consequences, in terms of sediment yield and phosphorus export, of hazelnut tree cultivation were estimated on different areas of the basin with and without the application of a best management practice (BMP) that consists of growing meadow under the trees. The GLEAMS results were successively extended to basin scale thanks to the application of a purposely designed regression model and of a GIS. The main conclusions can be summarized as follows: The effectiveness of the above-mentioned BMP is always greater for erosion reduction than for particulate P reduction, whatever the slope value considered; moreover, the effectiveness with reference to both particulate P and sediment yield production decreases as the slope increases. The proposed approach, being completely distributed, represents a considerable step ahead compared to the semidistributed or lumped approaches, which are traditionally employed in research into tools to support the decision-making process for land-use planning aimed at water pollution control.     

ANN OUTPUT UPDATING OF LUMPED CONCEPTUAL RAINFALL/RUNOFF FORECASTING MODELS1

ABSTRACT: Artificial neural networks (ANNs) are tested for the output updating of one‐day‐ahead and three‐day‐ahead streamflow forecasts derived from three lumped conceptual rainfall/runoff (R‐R) models: the GR4J, the IHAC, and the TOPMO. ANN output updating proved superior to a parameter updating scheme and to the ‘simple’ output updating scheme, which always replicates the last observed forecast error. In fact, ANN output updating was able to compensate for large differences in the initial performance of the three tested lumped conceptual R‐R models, which the other tested updating approaches were not able to achieve. This is done mainly by incorporating input vectors usually exploited for ANN R‐R modeling such as previous rainfall and streamflow observations, in addition to the previous observed error. For one‐day‐ahead forecasts, the performance of all three lumped conceptual R‐R models, used in conjunction with ANN output updating, was equivalent to that of the ANN R‐R model. For three‐day‐ahead forecasts, the performance of the ANN‐output‐updated conceptual models was even superior to that of the ANN R‐R model, revealing that the conceptual models are probably performing some tasks that the ANN R‐R model cannot map. However, further testing is needed to substantiate the last statement.