Uncertainty In Model Simulations Research Articles

Enhanced SWAT calibration through intelligent range-based parameter optimization

Hydrological models are vital tools in environmental management. Weaknesses in model robustness for hydrological parameters transfer uncertainties to the model outputs. For streamflow, the optimized parameters are the primary source of uncertainty. A reliable calibration approach that reduces prediction uncertainty in model simulations is crucial for enhancing model robustness and reliability. The optimization of parameter ranges is a key aspect of parameter calibration, yet there is a lack of literature addressing the optimization of parameter ranges in hydrological models. In this paper, we introduce a parameter calibration strategy that applies a clustering technique, specifically the Self-Organizing Map (SM), to intelligently navigate the parameter space during the calibration of the Soil and Water Assessment Tool (SWAT) model for monthly streamflow simulation in the Baishan Basin, Jilin Province, China. We selected the representative algorithm, the Sequential Uncertainty Fitting version 2 (SUFI-2), from the commonly used SWAT Calibration and Uncertainty Programs for comparison. We developed three schemes: SUFI-2, SUFI-2-Narrowing Down (SUFI-2-ND), and SM. Multiple diagnostic error metrics were used to compare simulation accuracy and prediction uncertainty. Among all schemes, SM outperformed the others in describing watershed streamflow, particularly excelling in the simulation of spring snowmelt runoff (baseflow period). Additionally, the prediction uncertainty was effectively controlled, demonstrating the SM's adaptability and reliability in the interval optimization process. This provides managers with more credible prediction results, highlighting its potential as a valuable calibration tool in hydrological modeling.

Implication of community-level ecophysiological parameterization to modelling ecosystem productivity: a case study across nine contrasting forest sites in eastern China

Parameterization is a critical step in modelling ecosystem dynamics. However, assigning parameter values can be a technical challenge for structurally complex natural plant communities; uncertainties in model simulations often arise from inappropriate model parameterization. Here we compared five methods for defining community-level specific leaf area (SLA) and leaf C:N across nine contrasting forest sites along the North–South Transect of Eastern China, including biomass-weighted average for the entire plant community (AP_BW) and four simplified selective sampling (biomass-weighted average over five dominant tree species [5DT_BW], basal area weighted average over five dominant tree species [5DT_AW], biomass-weighted average over all tree species [AT_BW] and basal area weighted average over all tree species [AT_AW]). We found that the default values for SLA and leaf C:N embedded in the Biome-BGC v4.2 were higher than the five computational methods produced across the nine sites, with deviations ranging from 28.0 to 73.3%. In addition, there were only slight deviations (< 10%) between the whole plant community sampling (AP_BW) predicted NPP and the four simplified selective sampling methods, and no significant difference between the predictions of AT_BW and AP_BW except the Shennongjia site. The findings in this study highlights the critical importance of computational strategies for community-level parameterization in ecosystem process modelling, and will support the choice of parameterization methods.

Combining CloudSat/CALIPSO and MODIS measurements to reconstruct tropical convective cloud structure

Tropical convective clouds are a crucial component of the Earth's weather and climate system. However, there are still large uncertainties in model simulations of tropical convective clouds and precipitation. In this study, CloudSat/CALIPSO and MODIS measurements are combined to provide vertical convective cloud structures under horizontal storm context and structures of convective cores. Convective cells are identified with MODIS brightness temperature at 11 μm (BT11) and convective centers are determined for each cell by finding the minimum BT11 (BT11min). The difference between BT11 and BT11min is calculated to provide the distance to the convective center as in brightness temperature. Therefore, detailed convective cloud structures from centers to the stratiform regions could be shown as in the contoured frequency by temperature diagram (CFTD) of radar reflectivity factor (Ze). A group of strongly attenuated Ze profiles is identified near the convective center in Africa during daytime, and the occurrence of this group of profiles is much smaller in the west Pacific during daytime. The out-of-phase day/night variations of convective system structures over ocean and over land are also analyzed. It is shown that over ocean regions, rather than land regions, nighttime convective clouds have larger Ze than daytime ones between 210 K to 260 K, but they have smaller Ze at temperatures colder than 210 K near the cloud top. The convective cores are analyzed by separating and classifying those strongly attenuated Ze profiles as group one convective cores representing the developing stage, and the rest regular profiles as group two convective cores representing the later stage. As BT11min decreases for group one, the cores grow stronger with the maximum Ze shifts into lower temperature; and as BT11min decreases for group two, the CFTD for the cores gradually shifts to larger Ze for temperatures colder than 273.15 K and smaller Ze for temperatures warmer than 273.15 K. These observed composite structures of convective clouds and land/ocean and day/night variations provide crucial information for model evaluations.

Does soil thinning change soil erodibility? An exploration of long-term erosion feedback systems

Abstract. Soil erosion rates on arable land frequently exceed the pace at which new soil is formed. This imbalance leads to soil thinning (i.e. truncation), whereby subsoil horizons and their underlying parent material become progressively closer to the land surface. As soil erosion is a selective process and subsurface horizons often have contrasting properties to the original topsoil, truncation-induced changes to soil properties might affect erosion rates and runoff formation through a soil erosion feedback system. However, the potential interactions between soil erosion and soil truncation are poorly understood due to a lack of empirical data and the neglection of long-term erodibility dynamics in erosion simulation models. Here, we present a novel model-based exploration of the soil erosion feedback system over a period of 500 years using measured soil properties from a diversified database of 265 agricultural soil profiles in the UK. For this, we adapted the Modified Morgan–Morgan–Finney model (MMMF) to perform a modelling experiment in which topography, climate, land cover, and crop management parameters were held constant throughout the simulation period. As selective soil erosion processes removed topsoil layers, the model gradually mixed subsurface soil horizons into a 0.2 m plough layer and updated soil properties using mass-balance mixing models. Further, we estimated the uncertainty in model simulations with a forward error assessment. We found that modelled erosion rates in 99 % of the soil profiles were sensitive to truncation-induced changes in soil properties. The soil losses in all except one of the truncation-sensitive profiles displayed a decelerating trend, which depicted an exponential decay in erosion rates over the simulation period. This was largely explained by decreasing silt contents in the soil surface due to selective removal of this more erodible particle size fraction and the presence of clayey or sandy substrata. Moreover, the soil profiles displayed an increased residual stone cover, which armoured the land surface and reduced soil detachment. Contrastingly, the soils with siltier subsurface horizons continuously replenished the plough layer with readily erodible material, which prevented the decline of soil loss rates over time. Although our results are limited by the edaphoclimatic conditions represented in our data, as by our modelling assumptions, we have demonstrated how modelled soil losses can be sensitive to erosion-induced changes in soil properties. These findings are likely to affect how we calculate soil lifespans and make long-term projections of land degradation.

Probabilistic Assessment of Cereal Rye Cover Crop Impacts on Regional Crop Yield and Soil Carbon

Field research for exploring the impact of winter cover crops (WCCs) integration into cropping systems is resource intensive, time-consuming and offers limited application beyond the study area. To bridge this gap, we used the APSIM model, to simulate corn (Zea mays L.)-rye (Secale cereale L.)-corn-rye and corn-rye-soybean (Glycine max L.)-rye rotations in comparison with corn-corn and corn-soybean rotations across the state of Illinois at a spatial resolution of 5 km × 5 km from 2000 to 2020 to study the impact of WCCs on soil organic carbon (SOC) dynamics and crop production. By propagating the uncertainty in model simulations associated with initial conditions, weather, soil, and management practices, we estimated the probability and the expected value of change in crop yield and SOC following WCC integration. Our results suggest that integrating cereal rye into the crop rotations imparted greater yield stability for corn across the state. It was found that the areas with low probability of increase in SOC (p &lt; 0.75) responded equally well for soil carbon sequestration through long term adoption of WCCs. This study presents the most complete uncertainty accounting of WCC benefits across a broad region and provides greater insights into the spatiotemporal variability of WCCs benefits for increasing WCC adoption rate.

Water Balance Uncertainty of a Hydrologic Model to Lengthy Drought and Storm Events in Managed Forest Catchments, Eastern Australia

Interest has grown in applying hydrologic models in managed catchments despite uncertainties around model inputs and empirical relationships to simulate complex geo-hydrological processes of streamflow and sediment variations. Unquantified interactions between geophysical, climate and management indices can also increase simulation uncertainties. Calibration of model outputs against observed values allows identification of the most influential variables and their optimised ranges by which model performance can be enhanced. A rainfall-runoff Soil and Water Assessment Tool (SWAT) model was utilised for four catchments in northern New South Wales, Australia to simulate time series of streamflow across varying rainfall regimes, from dry seasons from 2002 towards rainy 2009. Parameters causing a substantial change on model streamflow outputs were first identified using a sensitivity analysis which indicated that hydrologic factors governing the sources of water supply were critical parameters. These sensitive variables were substantially derived from groundwater modules, basic flow in the main channel, and management practices. Statistical tests of between-catchment differences showed that model simulations performed better in a catchment where the sole rain gauge was installed, while also having the narrowest variations in simulated values (r-index = 0.02). In contrast, the highest uncertainty of model simulations was found in the furthest catchment from the rain gauge where there was not a satisfactory agreement with observed data. Yearly differences between 2002 and 2009 indicated an overestimation of streamflow during low flow periods. However, the calibration process performed well in most peak flows where estimations followed the respective observed values. Long-term dry periods between 2002 and 2007 resulted in an overestimated baseflow by predicting an unrealistic recharge infiltrating aquifers.

Estimated regional CO2flux and uncertainty based on an ensemble of atmospheric CO2inversions

Abstract. Global and regional sources and sinks of carbon across the earth's surface have been studied extensively using atmospheric carbon dioxide (CO2) observations and atmospheric chemistry-transport model (ACTM) simulations (top-down/inversion method). However, the uncertainties in the regional flux distributions remain unconstrained due to the lack of high-quality measurements, uncertainties in model simulations, and representation of data and flux errors in the inversion systems. Here, we assess the representation of data and flux errors using a suite of 16 inversion cases derived from a single transport model (MIROC4-ACTM) but different sets of a priori (bottom-up) terrestrial biosphere and oceanic fluxes, as well as prior flux and observational data uncertainties (50 sites) to estimate CO2 fluxes for 84 regions over the period 2000–2020. The inversion ensembles provide a mean flux field that is consistent with the global CO2 growth rate, land and ocean sink partitioning of −2.9 ± 0.3 (± 1σ uncertainty on the ensemble mean) and −1.6 ± 0.2 PgC yr−1, respectively, for the period 2011–2020 (without riverine export correction), offsetting about 22 %–33 % and 16 %–18 % of global fossil fuel CO2 emissions. The rivers carry about 0.6 PgC yr−1 of land sink into the deep ocean, and thus the effective land and ocean partitioning is −2.3 ± 0.3 and −2.2 ± 0.3, respectively. Aggregated fluxes for 15 land regions compare reasonably well with the best estimations for the 2000s (∼ 2000–2009), given by the REgional Carbon Cycle Assessment and Processes (RECCAP), and all regions appeared as a carbon sink over 2011–2020. Interannual variability and seasonal cycle in CO2 fluxes are more consistently derived for two distinct prior fluxes when a greater degree of freedom (increased prior flux uncertainty) is given to the inversion system. We have further evaluated the inversion fluxes using meridional CO2 distributions from independent (not used in the inversions) aircraft and surface measurements, suggesting that the ensemble mean flux (model–observation mean ± 1σ standard deviation = −0.3 ± 3 ppm) is best suited for global and regional CO2 flux budgets than an individual inversion (model–observation 1σ standard deviation = −0.35 ± 3.3 ppm). Using the ensemble mean fluxes and uncertainties for 15 land and 11 ocean regions at 5-year intervals, we show promise in the capability to track flux changes toward supporting the ongoing and future CO2 emission mitigation policies.

Attribution of NDVI Dynamics over the Globe from 1982 to 2015

Satellite remote sensing has witnessed a global widespread vegetation greening since the 1980s. However, reliable observation-based quantitative knowledge on global greening remains obscure due to uncertainties in model simulations and the contribution of natural variability is largely unknown. Here, we revisit the attribution of global vegetation changes using the Time Series Segment and Residual Trend (TSS-RESTREND) method. Results showed global vegetation significantly greening over 40.6% of the vegetated grids, whereas vegetation significantly browning over 11.6% of the vegetated grids. The attribution results based on the TSS-RESTREND method show that CO2 fertilization (CO2) plays an influential role in vegetation changes over 61.4% of the global vegetated areas, followed by land use (LU, 23.5%), climate change (CC, 7.3%), and climate variability (CV, 1.5%). The vegetation greening can be largely attributed to CO2 fertilization while the vegetation browning is mainly caused by LU. Meanwhile, we also identify positive impacts of LU and CC on vegetation change in arid regions but negative impacts in humid regions. Our findings indicate spatial heterogeneity in causes behind global vegetation changes, providing more detailed references for global vegetation modeling.

Future evolution of global land surface air temperature trend based on Coupled Model Intercomparison Project Phase 6 models

AbstractIn recent decades, global warming has been an indisputable fact. The increase in extreme events accompanied by the rise in temperature has huge influences on many aspects of the society and economy globally. As a result, forecasting how global temperatures will change and evolve in the future is crucial, but it is restricted by the accuracy and uncertainty of model simulations. Based on ensemble empirical mode decomposition (EEMD) method, this study first evaluates the performance of the CMIP6 models in simulating different timescale components including high‐frequency components (HFC), low‐frequency components (LFC), and secular trend (ST) of surface air temperature (SAT). The results show that the performance of the CMIP6 models in simulating the ST is better. Following the observation constraint method's correction of CMIP6 model simulations, we investigate the evolution characteristics of future temperature secular trends under different Shared Socioeconomic Pathway (SSP) scenarios. The results show that under SSP245 scenario, the rise of SAT first appeared in northwest and east North America and eastern Europe. The spatial distribution of evolution characteristic under the SSP585 scenario is similar to that of the SSP245 scenario but with a larger magnitude. However, more pronounced differences appear in warming rate, which shows a downward trend over time under the SSP245 scenario and an upward trend with time under the SSP585 scenario.

An Introduction to the E3SM Special Collection: Goals, Science Drivers, Development, and Analysis

AbstractSupported by the U.S. Department of Energy (DOE), the Energy Exascale Earth System Model (E3SM) project aims to optimize the use of DOE resources to address the grand challenge of actionable predictions of Earth system variability and change. This requires sustained advancement to (1) integrate model development with leading‐edge computational advances toward ultra‐high‐resolution modeling; (2) represent the coupled human‐Earth system to address energy sector vulnerability to variability and change; and (3) address uncertainty in model simulations and projections. Scientific development of the E3SM modeling system is driven by the simulation requirements in three overarching science areas centering on understanding the Earth's water cycle, biogeochemistry, and cryosphere systems and their future changes. This paper serves as an introduction to the E3SM special collection, which includes 50 papers published in several AGU journals. It provides an overview of the E3SM project, including its goals and science drivers. It also provides a brief history of the development of E3SM version 1 and highlights some key findings from papers included in the special collection.

Advancing model calibration and uncertainty analysis of SWAT models using cloud computing infrastructure: LCC-SWAT

Abstract Calibration and uncertainty analysis of a complex, over-parameterized environmental model such as the Soil and Water Assessment Tool (SWAT) requires thousands of simulation runs and multiple calibration iterations. A parallel calibration system is thus desired that can be deployed on cloud-based architectures for reducing calibration runtime. This paper presents a cloud-based calibration and uncertainty analysis system called LCC-SWAT that is designed for SWAT models. Two optimization techniques, sequential uncertainty fitting (SUFI-2) and dynamically dimensioned search (DDS), have been implemented in LCC-SWAT. Moreover, the cloud-based system has been deployed on the Southern Ontario Smart Computing Innovation Platform's (SOSCIP) Cloud Analytics platform for diagnostic assessment of parallel calibration runtime on both single-node and multi-node CPU architectures. Unlike other calibrations/uncertainty analysis systems developed on the cloud, this system is capable of generating a comprehensive set of statistical information automatically, which facilitates broader analyses of the performance of the SWAT models. Experimental results on SWAT models of different complexities showed that LCC-SWAT can reduce runtime significantly. The runtime reduction is more pronounced for more complex and computationally intensive models. However, the reported runtime efficiency is significantly higher for single node systems. Comparative experiments with DDS and SUFI-2 show that parallel DDS outperforms parallel SUFI-2 in terms of both parameter identifiability and reducing uncertainty in model simulations. LCC-SWAT is a flexible calibration system and other optimization algorithms and asynchronous parallelization strategies can be added to it in future.

A framework for determining the maximum allowable external load that will meet a guarantee probability of achieving water quality targets

Quantifying the maximum external pollutant loading is extremely important for environmental management and ecological restoration. However, huge uncertainty exists in the process of determining accurate external pollutant loads discharging into surface water bodies (e.g., rivers, reservoirs, and bays). In this paper, a comprehensive framework is proposed for determining the maximum allowable external load by combining a dynamic nutrient-balance model with the guarantee probability of achieving a specific water quality target. As an important drinking water source for Beijing, the Miyun Reservoir was chosen as a case study because it is experiencing increasing eutrophication. The main results are as follows. ① The nutrient-balance model has shown a good fit to field observations both in calibration and validation periods using the modified Generalized Likelihood Uncertainty Estimation (GLUE). ② Feasible concentration targets were determined for total phosphorus (TP), total nitrogen (TN), and chlorophyll-a as 0.01 mg/L, 0.76 mg/L, and 4.91 μg/L, respectively. ③ The allowable external load of TP is estimated as 45.10–54.14 t, 23.76–29.58 t, and 8.30–12.78 t for guarantee probabilities of TP control target (e.g., 0.01 mg/L) of 25, 50, and 70%, respectively. While the external TN flux should be reduced by 200.21–480.73 t, 429.33–764.45 t, and 642.40–1069.59 t to meet the TN control target (e.g., 0.76 mg/L) at 25, 50, and 70% guarantee probabilities, respectively, The wide range of allowable external nutrient loading reflects the 95% confidence intervals of the load reduction analysis and indicates the importance of model simulation uncertainty and interpretation of the water quality objective. This paper provides a scientifically sound approach to water quality maintenance for the Miyun Reservoir and other surface water bodies.

Clumped isotope constraints on changes in latest Pleistocene hydroclimate in the northwestern Great Basin: Lake Surprise, California

AbstractDuring the Last Glacial Maximum (LGM) and subsequent deglaciation, the Great Basin in the southwestern United States was covered by numerous extensive closed-basin lakes, in stark contrast with the predominately arid climate observed today. This transition from lakes in the Late Pleistocene to modern aridity implies large changes in the regional water balance. Whether these changes were driven by increased precipitation rates due to changes in atmospheric dynamics, decreased evaporation rates resulting from temperature depression and summer insolation changes, or some combination of the two remains uncertain. The factors contributing to these large-scale changes in hydroclimate are critical to resolve, given that this region is poised to undergo future anthropogenic-forced climate changes with large uncertainties in model simulations for the 21st century. Furthermore, there are ambiguous constraints on the magnitude and even the sign of changes in key hydroclimate variables between the Last Glacial Maximum and the present day in both proxy reconstructions and climate model analyses of the region. Here we report thermodynamically derived estimates of changes in temperature, precipitation, and evaporation rates, as well as the isotopic composition of lake water, using clumped isotope data from an ancient lake in the northwestern Great Basin, Lake Surprise (California). Compared to modern climate, mean annual air temperature at Lake Surprise was 4.7 °C lower during the Last Glacial Maximum, with decreased evaporation rates and similar precipitation rates to modern. During the mid-deglacial period, the growth of Lake Surprise implied that the lake hydrologic budget briefly departed from steady state. Our reconstructions indicate that this growth took place rapidly, while the subsequent lake regression took place over several thousand years. Using models for precipitation and evaporation constrained from clumped isotope results, we determine that the disappearance of Lake Surprise coincided with a moderate increase in lake temperature, along with increasing evaporation rates outpacing increasing precipitation rates. Concomitant analysis of proxy data and climate model simulations for the Last Glacial Maximum are used to provide a robust means to understand past climate change, and by extension, predict how current hydroclimates may respond to expected future climate forcings. We suggest that an expansion of this analysis to more basins across a larger spatial scale could provide valuable insight into proposed climate forcings, and aid in climate model process depiction. Ultimately, our analysis highlights the importance of temperature-driven evaporation as a mechanism for lake growth and retreat in this region.

A bootstrap approach for the parameter uncertainty of an urban-specific rainfall-runoff model

The predictions made using rainfall-runoff models are inherently uncertain and it is important to recognize and account for this uncertainty, especially in urban watersheds due to the high flood risk in these areas. Recent studies on hydrological model uncertainty mostly refer to the identification of model parameter uncertainty. However, such studies are somewhat limited using the bootstrap approach, a nonparametric method which makes less prior assumptions on the model structure and thus is more flexible. Hence, a residual-based bootstrap approach associated with the SCE-UA global optimization algorithm is demonstrated in this study for the analysis of calibrated parameter uncertainty and its subsequent effect on the model simulation of an urban-specific rainfall-runoff model, urban storage function (USF) model, under two different data scenarios of individual event-based and whole data-based scenarios. Initially, the parameter uncertainty was expressed by estimating the confidence interval (CI) of the USF model parameters obtained from bootstrapping and then the parameters from the highest to the lowest uncertainties were derived by utilizing two newly proposed parameter uncertainty indices which can make the best use of CI. Moreover, investigations on the effect of calibrated parameter uncertainty on model simulations revealed that the model was able to bracket most of the observations within the prediction range of considered scenarios. This further indicates that the residual-based bootstrap approach along with the SCE-UA method reasonably well predicted the uncertainty range of the USF model. For a better understanding of simulation uncertainty, we defined and demonstrated two model simulation uncertainty indices and these indices could be useful in future studies to analyze the simulation uncertainty of different rainfall-runoff models in the watersheds worldwide.

Influences of planetary boundary layer mixing parameterization on summertime surface ozone concentration and dry deposition over North China

We present a regional modeling study that analyzes how planetary boundary layer (PBL) and surface layer parameterizations influence surface ozone concentrations and dry deposition fluxes over the Beijing-Tianjin-Hebei region in summer (July 2015). We use the Weather Research and Forecasting Model coupled to Chemistry (WRF-Chem) to simulate surface ozone concentration and dry deposition, and examine three PBL schemes: the Yonsei University (YSU), Mellor–Yamada–Janjić (MYJ), and Asymmetric Convective Model version 2 (ACM2) schemes. The model sensitivity to surface layer schemes is also tested by coupling the ACM2 PBL scheme with either the revised MM5-similarity scheme or the Pleim-Xiu surface layer scheme. Key physical and chemical factors for ozone dry deposition parameterization are analyzed to explore the root causes of model discrepancies. We find that all simulations overestimate the summertime daily mean ozone concentrations over North China (42 ppbv in observations vs. 43–50 ppbv in model results for July 2015) that are caused by high biases in daytime ozone and partly compensated by low biases in nighttime ozone. The YSU scheme has the largest overestimate in daily mean ozone concentration, but best reproduces the ozone diurnal cycle. The ACM2 scheme shows the largest underestimates of surface ozone over North China during nighttime, which can be explained by its weakest vertical mixing leading to high NOx concentrations and strong ozone titration near surface. The choices of PBL and surface layer schemes lead to over 20% differences in ozone dry deposition fluxes due to differences in simulated surface ozone concentrations and dry deposition velocities. We find the differences in dry deposition velocity are mainly caused by differences in Monin-Obukhov length during nighttime and surface temperature during daytime. Our study emphasizes the needs to better understand these key PBL and surface factors for reducing the uncertainties in model simulation of surface ozone concentration and dry deposition.

Convective Entrainment Rates Estimated From Aura CO and CloudSat/CALIPSO Observations and Comparison With GEOS‐5

AbstractEntrainment rate (λ) in convective parameterizations remains a sensitive parameter with much uncertainty in model simulations. This study estimates λ using carbon monoxide (CO) measurements jointly from the Microwave Limb Sounder and Tropospheric Emission Spectrometer onboard the Aura satellite, associated with deep convective cases identified by CloudSat and CALIPSO observations. CO is treated as a conserved quantity over convective transport time scales, and a simple entraining‐plume model is used to derive entrainment rates. The relationships of the observational estimates of λ as a function of convective height, environmental relative humidity, and convective available potential energy (CAPE) derived from Atmospheric Infrared Sounder data are compared with those from Goddard Earth Observing System Model (GEOS‐5) simulations. Bulk statistics of λ show that the values of λ are predominately below 20%/km for deep convection and the occurrence frequency of any λ decreases with increasing λ. Composite λ values are generally lower in the tropics compared to northern midlatitudes in both observations and the GEOS‐5 model. A decrease of λ with increasing convective height is found in both observations and model simulations. We also find that λ tends to decrease with increasing CAPE in the observation‐based λs and plume‐based GEOS‐5 λs, although the model given λs have a nonmonotonic relation with CAPE. The observed λs have a weak relation with lower‐to‐middle tropospheric relative humidity, while both GEOS‐5 plume‐based and given λ increase with increasing relative humidity.

How Important Are Model Structural and Contextual Uncertainties when Estimating the Optimized Performance of Water Resource Systems?

AbstractUncertainty in simulating water resource systems (WRSs) makes it difficult to assess how effective different water management decisions will be. Uncertainty in simulation models can undermine the credibility of simulation and optimization studies and the uptake of their results. We identify different sources of uncertainty in WRS models and find that structural uncertainty (i.e., around definition of interrelationships within the system) and contextual uncertainty (i.e., around definition of the system boundaries) are rarely considered when simulating and optimizing WRSs. We propose a methodology to quantify the effects of structural and contextual uncertainties on the estimated performance of optimized water management decisions and demonstrate that they have a significant impact on a real‐world case study of a pumped‐storage system in the UK. To the best of the authors' knowledge, this is the first study to consider the impact of these types of uncertainty on optimized operating policies and their simulated performances. Our main finding is that of all the considered uncertainties, the assumptions made about context—specifically around the level of cooperation between neighboring water companies—had the greatest impact on performance estimates. This is important because few WRSs exist in isolation, yet discussion of the effects that a given definition of the system boundaries have on the simulation/optimization results is uncommon. We also highlight the significance of adequately considering aleatory uncertainty when evaluating performance estimates—something that few studies do—and present a simple technique to justify the sample size used for the evaluation of optimization results.

Modeling and uncertainty analysis of seawater intrusion in coastal aquifers using a surrogate model: a case study in Longkou, China

Seawater intrusion is a complex problem involving groundwater with variable density. It has an important impact on the development and utilization of groundwater resources in coastal areas. In previous studies, numerical simulations of seawater intrusion were performed by assigning fixed values to model parameters, thereby not taking into account the influence of stochastic variability of hydrogeological parameters on model predictions. In this study, we developed a three-dimensional mathematical model of seawater intrusion by modeling vertical solute transport due to density changes resulting from seawater intrusion. We then applied a Monte Carlo–based method to analyze uncertainty in model simulations of seawater intrusion. Because the Monte Carlo method requires the simulation model to be repeatedly run, we used the Kriging method to build a surrogate model that significantly reduced the computational load compared to the simulation model while ensuring the desired accuracy. Then we used a local sensitivity analysis method to select the two parameters with the greatest influence on model output. We treated the two selected parameters as random variables, and results show that (1) the three-dimensional, variable-density seawater intrusion model can effectively simulate and predict the distribution of seawater intrusion in the study area; (2) the local sensitivity analysis can accurately identify the hydrogeological parameters that most influence model output; (3) the uncertainty analysis based on the surrogate model reduces computing time substantially and provides a realistic assessment of the effect of hydrogeological parameter variability on seawater intrusion numerical simulation results.

Assessment of atmospheric aerosols from two reanalysis products over Australia

Assessments of atmospheric aerosols from reanalysis are important for understanding uncertainty in model simulations, and ultimately predictions, such as for solar power or air quality forecasts and assessments. This study intercompares total aerosol optical depth (AOD) and dust AOD (DAOD) from two global reanalyses datasets, the European Centre for Medium-Range Weather Forecasts (ECMWF) Monitoring Atmospheric Composition and Climate (MACC) and the NASA Modern-Era Retrospective Analysis for Research-2 (MERRA-2). These are evaluated against AeroSpan (Aerosol characterisation via Sun photometry: Australian Network) ground observations which forms part of the Aerosol Robotic Network (AERONET) over the Australian continent for the 2002–2012 period. During dust storms, AeroSpan/AERONET AOD measurements were missing due to cloud screening. To overcome validation limitations in sun photometry for dust events, a nephelometer's scattering coefficient is qualitatively compared against reanalysis of DAOD at a key dust storm activation site, Tinga Tingana in South Australia (~200 km east of Lake Eyre). A specific extreme event that occurred in 2009 originating from the Lake Eyre basin, a major dust source covering one-sixth of Australia, was studied. The results show that MERRA-2 reanalysis overestimates monthly total AOD twice as much compared to AeroSpan/AERONET ground observations but seems better correlated against AeroSpan/AERONET than ECMWF/MACC. Mean data of MERRA-2 time series over 10 years provide lower DAOD values and lower dust aerosol estimates than ECMWF/MACC reanalysis (over the Lake Eyre basin with spatial averaging). Specifically at Tinga Tingana, the correlation from MERRA-2 (0.45 correlation) and ECMWF/MACC (0.43 correlation) against AeroSpan/AERONET's AOD were similar. Between MERRA-2 and ECMWF/MACC decade long daily gridded DAOD, the correlation coefficient was high at 0.73, again indicating similarity between the datasets. MERRA-2 total AOD correlation is significantly higher (by 0.26) against AeroSpan/AERONET than ECMWF/MACC. MERRA-2 also provides higher AOD values in extreme cases which may correspond to dust storms. During dust storms, a hybrid strategy using nephelometers and hourly reanalysis from MERRA-2 is able to identify dust storms better than AeroSpan/AERONET. Overall, this work can enable and inform better aerosol data assimilation into forecast models such as for solar energy, agriculture or air quality over Australia.

Impact of robustness of hydrological model parameters on flood prediction uncertainty

The robustness of hydrological model parameter values in flood predictions is a known area of concern, but there is a lack of a comprehensive approach to the handling of model parameter robustness, model simulation uncertainty and multiobjective model calibration when calibrating multiple flood data sets. For investigation of the impact of robustness of hydrological model parameters on flood simulation uncertainty, this paper develops a Minimax‐Regret robust multiobjective optimisation framework for robust hydrological model parameter calibration and uncertainty analysis. The robustness is considered as an objective function in this study instead of a constraint as in previous research. A physically based semi‐distributed hydrological model is employed to illustrate the proposed framework in a midscale catchment. Results show that the proposed framework can effectively explore robust hydrological model parameter values and quantify flood simulation uncertainty. A trade‐off between robustness and Nash–Sutcliffe efficiency is found, implying that the better the Nash–Sutcliffe efficiency, the less robust the non‐dominated parameter values and that improving robustness alone cannot guarantee narrower uncertainty intervals and greater containing ratios. These results reveal that robustness should not be used alone to select behavioural parameter sets, and a balance has to be made between robustness and Nash–Sutcliffe efficiency.