Advanced Weather Generator Research Articles

Stochastic Hybrid Event Based and Continuous Approach to Derive Flood Frequency Curve

This study proposes a methodology that combines the advantages of the event-based and continuous models, for the derivation of the maximum flow and maximum hydrograph volume frequency curves, by combining a stochastic continuous weather generator (the advanced weather generator, abbreviated as AWE-GEN) with a fully distributed physically based hydrological model (the TIN-based real-time integrated basin simulator, abbreviated as tRIBS) that runs both event-based and continuous simulation. The methodology is applied to Peacheater Creek, a 64 km2 basin located in Oklahoma, United States. First, a continuous set of 5000 years’ hourly weather forcing series is generated using the stochastic weather generator AWE-GEN. Second, a hydrological continuous simulation of 50 years of the climate series is generated with the hydrological model tRIBS. Simultaneously, the separation of storm events is performed by applying the exponential method to the 5000- and 50-years climate series. From the continuous simulation of 50 years, the mean soil moisture in the top 10 cm (MSM10) of the soil layer of the basin at an hourly time step is extracted. Afterwards, from the times series of hourly MSM10, the values associated to all the storm events within the 50 years of hourly weather series are extracted. Therefore, each storm event has an initial soil moisture value associated (MSM10Event). Thus, the probability distribution of MSM10Event for each month of the year is obtained. Third, the five major events of each of the 5000 years in terms of total depth are simulated in an event-based framework in tRIBS, assigning an initial moisture state value for the basin using a Monte Carlo framework. Finally, the maximum annual hydrographs are obtained in terms of maximum peak-flow and volume, and the associated frequency curves are derived. To validate the method, the results obtained by the hybrid method are compared to those obtained by deriving the flood frequency curves from the continuous simulation of 5000 years, analyzing the maximum annual peak-flow and maximum annual volume, and the dependence between the peak-flow and volume. Independence between rainfall events and prior hydrological soil moisture conditions has been proved. The proposed hybrid method can reproduce the univariate flood frequency curves with a good agreement to those obtained by the continuous simulation. The maximum annual peak-flow frequency curve is obtained with a Nash–Sutcliffe coefficient of 0.98, whereas the maximum annual volume frequency curve is obtained with a Nash–Sutcliffe value of 0.97. The proposed hybrid method permits to generate hydrological forcing by using a fully distributed physically based model but reducing the computation times on the order from months to hours.

Can we replace observed forcing with weather generator in land surface modeling? Insights from long-term simulations at two contrasting boreal sites

This study evaluates the simulation of water balance components at half-hourly time steps from the Canadian Land Surface Scheme (CLASS) when driven by a 500-year stochastic meteorological data set produced by the Advanced WEather GENerator (AWE-GEN) at two boreal sites with contrasting water availability. The CLASS was driven by ERA5 reanalysis data (CLASS-CTL) over 39 years and its output was used as a surrogate for land surface observations. At both sites, the mean monthly and annual values of all meteorological variables used to drive CLASS, including precipitation, are well captured by AWE-GEN, but their variability is, sometimes, biased. In general, CLASS driven by stochastic data (CLASS-WG) tends to produce higher evapotranspiration compared to values simulated by CLASS-CTL, especially during spring and summer at the wet site. The interannual evapotranspiration-precipitation and runoff-precipitation relationships derived from CLASS-WG and those derived from CLASS-CTL were very similar to each other at the dry site; they both indicate that evapotranspiration and runoff are limited by water availability. At the wet site, however, CLASS-WG only captured well the interannual runoff-precipitation relationship. The sensitivity analysis shows that CLASS water fluxes are particularly affected by the replacement of physically consistent input time series of incoming short-wave radiation, precipitation, temperature, and specific humidity. In conclusion, the results show that even though a weather generator can produce coherent climate time series, the use of this synthetic data as meteorological forcing in a physically based land surface model does not necessarily reproduce the complex surface processes, such as the surface water fluxes. More studies are encouraged to further analyze the constraints of this framework.

Dependence Between Extreme Rainfall Events and the Seasonality and Bivariate Properties of Floods. A Continuous Distributed Physically-Based Approach

This paper focuses on proposing the minimum number of storms necessary to derive the extreme flood hydrographs accurately through event-based modelling. To do so, we analyzed the results obtained by coupling a continuous stochastic weather generator (the Advanced WEather GENerator) with a continuous distributed physically-based hydrological model (the TIN-based real-time integrated basin simulator), and by simulating 5000 years of hourly flow at the basin outlet. We modelled the outflows in a basin named Peacheater Creek located in Oklahoma, USA. Afterwards, we separated the independent rainfall events within the 5000 years of hourly weather forcing, and obtained the flood event associated to each storm from the continuous hourly flow. We ranked all the rainfall events within each year according to three criteria: Total depth, maximum intensity, and total duration. Finally, we compared the flood events obtained from the continuous simulation to those considering the N highest storm events per year according to the three criteria and by focusing on four different aspects: Magnitude and recurrence of the maximum annual peak-flow and volume, seasonality of floods, dependence among maximum peak-flows and volumes, and bivariate return periods. The main results are: (a) Considering the five largest total depth storms per year generates the maximum annual peak-flow and volume, with a probability of 94% and 99%, respectively and, for return periods higher than 50 years, the probability increases to 99% in both cases; (b) considering the five largest total depth storms per year the seasonality of flood is reproduced with an error of less than 4% and (c) bivariate properties between the peak-flow and volume are preserved, with an error on the estimation of the copula fitted of less than 2%.

Exploring stochastic climate uncertainty in space and time using a gridded hourly weather generator

Exploring the effects of climate change on the hydrological response at the local scale requires climate data at high spatial and temporal resolutions. This is best achieved by generating downscaled ensembles of future climate variables derived from climate models. For this purpose we present a methodology to re-parameterize the AWE-GEN-2d model (Advanced WEather GENerator for a two-dimensional grid). The model simulates key meteorological variables needed by hydrological models and is particularly suitable to explore the effects of stochastic (natural) climatic uncertainty, which is fundamental for hydrological applications, especially at sub-kilometer and hourly scales. Factors of change for different climate statistics are calculated from climate model simulations of present and future climates and subsequently applied to the statistics derived from observations to re-parameterize AWE-GEN-2d. The model abilities in generating an ensemble of future climate variables for the transient period 2020–2089 is presented with examples of precipitation and near-surface air temperature fields from hourly to multi-annual scales for a small mountainous region in the Swiss Alps. The stochastic uncertainty is examined for present and future periods and for spatial scales from the RCM scale (12-km, daily) to 2-km demonstrating the potential use of AWE-GEN-2d outputs. At the RCM scale, model results yield a small increase in annual precipitation (4%) which is within the stochastic uncertainty range for present and future periods (7%). At the fine scale of 2-km, the increase in annual precipitation can exceed the stochastic uncertainty, but for less than 10% of the domain area. On the contrary, changes in annual near-surface air temperature exceed stochastic uncertainty both at the RCM and finer scales. Stochastic climate uncertainty was concluded to be very similar when comparing present and future periods and 12-km and 2-km scales. The benefits of using AWE-GEN-2d in hydrological climate change impact assessments are finally discussed.

Stochastic Modeling of Rainfall Series in Kelantan Using an Advanced Weather Generator

Weather generator is a numerical tool that uses existing meteorological records to generate series of synthetic weather data. The AWE-GEN (Advanced Weather Generator) model has been successful in producing a broad range of temporal scale weather variables, ranging from the high-frequency hourly values to the low-frequency inter-annual variability. In Malaysia, AWE-GEN has produced reliable projections of extreme rainfall events for some parts of Peninsular Malaysia. This study focuses on the use of AWE-GEN model to assess rainfall distribution in Kelantan. Kelantan is situated on the north east of the Peninsular, a region which is highly susceptible to flood. Embedded within the AWE-GEN model is the Neyman Scott process which employs parameters to represent physical rainfall characteristics. The use of correct probability distributions to represent the parameters is imperative to allow reliable results to be produced. This study compares the performance of two probability distributions, Weibull and Gamma to represent rainfall intensity and the better distribution found was used subsequently to simulate hourly scaled rainfall series. Thirty years of hourly scaled meteorological data from two stations in Kelantan were used in model construction. Results indicate that both probability distributions are capable of replicating the rainfall series at both stations very well, however numerical evaluations suggested that Gamma performs better. Despite Gamma not being a heavy tailed distribution, it is able to replicate the key characteristics of rainfall series and particularly extreme values. The overall simulation results showed that the AWE-GEN model is capable of generating tropical rainfall series which could be beneficial in flood preparedness studies in areas vulnerable to flood.

Extreme Rainfall in the Eastern Region of Peninsular Malaysia

Modeling of daily rainfall data, particularly rainfall amount, is important not only for hydrological purposes but also for providing input for models of crop growth, design of urban drainage systems, land management systems and other environmental projects. This study tests the performance of Gamma and Weibull that are incorporated in the weather generator in simulating extreme rainfall and extreme dry/wet spell lengths for the eastern region of Peninsular Malaysia. Next, the monthly rainfall and the extreme rainfall are simulated using the best distribution. A weather generator known as Advanced Weather Generator (AWE-GEN) is chosen to model hourly rainfall time series in the eastern region of Peninsular Malaysia. Previously, Gamma distribution is incorporated in AWE-GEN to represent the rainfall intensity. This study proposes Weibull distribution and the performances between Gamma and Weibull distributions in generating rainfall series are compared based on Root Mean Square (RMSE) value error. Minimum RMSE value indicates the best distribution for eastern region. Results have shown that Gamma is the best distribution in simulating hourly rainfall in the east coast as compared to Weibull despite not being able to capture the peak/extreme values at several stations in the east coast of Peninsular Malaysia. The extreme wet spell length is well simulated although the extreme dry spell is under-estimated at all stations. Rainfall in November is contributing most of the annual rainfall followed by December. This study is relevant and very crucial for a tropical country like Malaysia which will benefit the policy makers in executing the strategic management of flooding in urban areas that is predominantly caused by convective activities.

Rainfall analysis in the northern region of Peninsular Malaysia

Modeling of rainfall is important for assessing the possible impacts of climate change. To achieve accurate projections of rainfall events, availability of sufficient hydrological station data is critical. Precipitation is one of the most important meteorological variables for hydrological modeling. In cases where long series of observed precipitation are not available, they can be stochastically generated by weather generators. Advanced Weather Generator (AWE-GEN) has been proven to generate precipitation data at the temperate climate regions with Gamma distribution being incorporated in the model to represent rainfall intensity. However, in a tropical climate such as Malaysia, some studies disputed the incorporation of Gamma distribution. As such, in this study, Weibull a heavy tail distribution is proposed to be used. The AWE-GEN has well performed in the wetter region such as the eastern of the peninsular. However, rainfall distribution within Peninsular Malaysia is highly variable temporally and spatially. The northern region is drier especially during the southwest monsoon season. This region receives minimal rain during the northeast monsoon due to the presence of the Titiwangsa Range which obstructs the region from getting rain by the north easterly winds. Therefore, the objectives of the study are two-fold. First, this study compares the performance of Gamma and Weibull that are incorporated in the AWE-GEN in simulating rainfall series for the northern region of the peninsular. Second, the monthly rainfall and the extreme rainfall series are simulated using the better distribution. The performances of Gamma and Weibull distributions are compared using the goodness of fit test, Root Mean Square Error (RMSE). Results showed that Gamma is the better distribution in simulating rainfall at rainfall stations located at the outer parts of the northern coast whereas Weibull is the better distribution for stations located in the interior parts of the northern coast. Hourly and daily extreme rainfalls seem to be well captured at all stations. Similarly, wet spell length is well simulated while in contrast, dry spell length is slightly underestimated at all stations. Overall, Gamma and Weibull produce commendable results in simulating extreme rainfall as well as wet spell length throughout the northern region of the peninsular. (c) 2017 The Authors. Published by IASE.

An advanced stochastic weather generator for simulating 2‐D high‐resolution climate variables

AbstractA new stochastic weather generator, Advanced WEather GENerator for a two‐dimensional grid (AWE‐GEN‐2d) is presented. The model combines physical and stochastic approaches to simulate key meteorological variables at high spatial and temporal resolution: 2 km × 2 km and 5 min for precipitation and cloud cover and 100 m × 100 m and 1 h for near‐surface air temperature, solar radiation, vapor pressure, atmospheric pressure, and near‐surface wind. The model requires spatially distributed data for the calibration process, which can nowadays be obtained by remote sensing devices (weather radar and satellites), reanalysis data sets and ground stations. AWE‐GEN‐2d is parsimonious in terms of computational demand and therefore is particularly suitable for studies where exploring internal climatic variability at multiple spatial and temporal scales is fundamental. Applications of the model include models of environmental systems, such as hydrological and geomorphological models, where high‐resolution spatial and temporal meteorological forcing is crucial. The weather generator was calibrated and validated for the Engelberg region, an area with complex topography in the Swiss Alps. Model test shows that the climate variables are generated by AWE‐GEN‐2d with a level of accuracy that is sufficient for many practical applications.

Climate Projections of Future Extreme Events in Malaysia

In Malaysia, extreme rainfall events are often linked to a number of environmental disasters such as landslides, monsoonal and flash floods. In response to the negative impacts of such disaster, studies assessing the changes and projections of extreme rainfall are vital in order to gather climate change information for better management of hydrological processes. This study investigates the changes and projections of extreme rainfall over Peninsular Malaysia for the period 2081-2100 based on the RCP 6.0 scenario. In particular, this study adopted the statistical downscaling method which enables high resolution, such as hourly data, to be used for the input. Short duration and high intensity convective rainfall is a normal feature of tropical rainfall especially in the western part of the peninsular. The proposed method, the Advanced Weather Generator model is constructed based on thirty years of hourly rainfall data from forty stations. To account for uncertainties, an ensemble multi-model of five General Circulation Model realizations is chosen to generate projections of extreme rainfall for the period 2081-2100. Results of the study indicate a possible increase in future extreme events for both the hourly and 24 h extreme rainfall with the latter showing a wider spatial distribution of increase.

Stochastic assessment of climate impacts on hydrology and geomorphology of semiarid headwater basins using a physically based model

AbstractHydrologic and geomorphic responses of watersheds to changes in climate are difficult to assess due to projection uncertainties and nonlinearity of the processes that are involved. Yet such assessments are increasingly needed and call for mechanistic approaches within a probabilistic framework. This study employs an integrated hydrology‐geomorphology model, the Triangulated Irregular Network‐based Real‐time Integrated Basin Simulator (tRIBS)‐Erosion, to analyze runoff and erosion sensitivity of seven semiarid headwater basins to projected climate conditions. The Advanced Weather Generator is used to produce two climate ensembles representative of the historic and future climate conditions for the Walnut Gulch Experimental Watershed located in the southwest U.S. The former ensemble incorporates the stochastic variability of the observed climate, while the latter includes the stochastic variability and the uncertainty of multimodel climate change projections. The ensembles are used as forcing for tRIBS‐Erosion that simulates runoff and sediment basin responses leading to probabilistic inferences of future changes. The results show that annual precipitation for the area is generally expected to decrease in the future, with lower hourly intensities and similar daily rates. The smaller hourly rainfall generally results in lower mean annual runoff. However, a non‐negligible probability of runoff increase in the future is identified, resulting from stochastic combinations of years with low and high runoff. On average, the magnitudes of mean and extreme events of sediment yield are expected to decrease with a very high probability. Importantly, the projected variability of annual sediment transport for the future conditions is comparable to that for the historic conditions, despite the fact that the former account for a much wider range of possible climate “alternatives.” This result demonstrates that the historic natural climate variability of sediment yield is already so high, that it is comparable to the variability for a projected and highly uncertain future. Additionally, changes in the scaling relationship between specific sediment yield/runoff and drainage basin area are detected.

Future projections of extreme precipitation using Advanced Weather Generator (AWE-GEN) over Peninsular Malaysia

Abstract. A stochastic downscaling methodology known as the Advanced Weather Generator, AWE-GEN, has been tested at four stations in Peninsular Malaysia using observations available from 1975 to 2005. The methodology involves a stochastic downscaling procedure based on a Bayesian approach. Climate statistics from a multi-model ensemble of General Circulation Model (GCM) outputs were calculated and factors of change were derived to produce the probability distribution functions (PDF). New parameters were obtained to project future climate time series. A multi-model ensemble was used in this study. The projections of extreme precipitation were based on the RCP 6.0 scenario (2081–2100). The model was able to simulate both hourly and 24-h extreme precipitation, as well as wet spell durations quite well for almost all regions. However, the performance of GCM models varies significantly in all regions showing high variability of monthly precipitation for both observed and future periods. The extreme precipitation for both hourly and 24-h seems to increase in future, while extreme of wet spells remain unchanged, up to the return periods of 10–40 years.

Interannual variability of evapotranspiration and vegetation productivity

AbstractInterannual variability of precipitation can influence components of the hydrological budget, affecting them directly and indirectly through adjustments in vegetation structure and function. We investigate the effects of fluctuations of annual precipitation on ecohydrological dynamics. Specifically, we use the advanced weather generator, AWE‐GEN, to simulate 200 years of hourly meteorological forcing obtained by imposing four types of precipitation annual process with identical long‐term mean. The generated time series force a mechanistic ecohydrological model, Tethys‐Chloris. Simulations with perturbed precipitation variability are performed for four locations characterized by different vegetation cover and climate. The results indicate that long‐term transpiration (T) and evapotranspiration (ET) fluxes as well as vegetation productivity expressed as Gross Primary Production (GPP) and Aboveground Net Primary Production (ANPP) are essentially unaffected by the imposed climate fluctuations. This finding supports the hypothesis of a relative insensitivity, except for water‐limited environments, of interannual evapotranspiration and vegetation productivity to annual climatic fluctuations, which are mostly reflected in the fluxes of deep leakage and runoff. The occurrence of short periods of favorable meteorological conditions randomly taking place within the year was found to be a better explanatory variable for interannual variability of ET and ANPP than average annual or growing season conditions. The results indicated that local, single‐site sensitivities are considerably smaller than those observed across climatic and vegetation spatial gradients and thus an important role of ecosystem reorganization in modifying ANPP and ET sensitivity in a changing climate is recognized.

Assessment of a stochastic downscaling methodology in generating an ensemble of hourly future climate time series

This study extends a stochastic downscaling methodology to generation of an ensemble of hourly time series of meteorological variables that express possible future climate conditions at a point-scale. The stochastic downscaling uses general circulation model (GCM) realizations and an hourly weather generator, the Advanced WEather GENerator (AWE-GEN). Marginal distributions of factors of change are computed for several climate statistics using a Bayesian methodology that can weight GCM realizations based on the model relative performance with respect to a historical climate and a degree of disagreement in projecting future conditions. A Monte Carlo technique is used to sample the factors of change from their respective marginal distributions. As a comparison with traditional approaches, factors of change are also estimated by averaging GCM realizations. With either approach, the derived factors of change are applied to the climate statistics inferred from historical observations to re-evaluate parameters of the weather generator. The re-parameterized generator yields hourly time series of meteorological variables that can be considered to be representative of future climate conditions. In this study, the time series are generated in an ensemble mode to fully reflect the uncertainty of GCM projections, climate stochasticity, as well as uncertainties of the downscaling procedure. Applications of the methodology in reproducing future climate conditions for the periods of 2000–2009, 2046–2065 and 2081–2100, using the period of 1962–1992 as the historical baseline are discussed for the location of Firenze (Italy). The inferences of the methodology for the period of 2000–2009 are tested against observations to assess reliability of the stochastic downscaling procedure in reproducing statistics of meteorological variables at different time scales.