Development of Boosted Machine Learning Models for Estimating Daily Reference Evapotranspiration and Comparison with Empirical Approaches

Proper irrigation scheduling and agricultural water management require a precise estimation of crop water requirement. In practice, reference evapotranspiration (ETo) is firstly estimated, and used further to calculate the evapotranspiration of each crop. In this study, two new coupled models were developed for estimating daily ETo. Two optimization algorithms, the shuffled frog-leaping algorithm (SFLA) and invasive weed optimization (IWO), were coupled on an adaptive neuro-fuzzy inference system (ANFIS) to develop and implement the two novel hybrid models (ANFIS-SFLA and ANFIS-IWO). Additionally, four empirical models with varying complexities, including Hargreaves–Samani, Romanenko, Priestley–Taylor, and Valiantzas, were used and compared with the developed hybrid models. The performance of all investigated models was evaluated using the ETo estimates with the FAO-56 recommended method as a benchmark, as well as multiple statistical indicators including root-mean-square error (RMSE), relative RMSE (RRMSE), mean absolute error (MAE), coefficient of determination (R2), and Nash–Sutcliffe efficiency (NSE). All models were tested in Tabriz and Shiraz, Iran as the two studied sites. Evaluation results showed that the developed coupled models yielded better results than the classic ANFIS, with the ANFIS-SFLA outperforming the ANFIS-IWO. Among empirical models, generally the Valiantzas model in its original and calibrated versions presented the best performance. In terms of model complexity (the number of predictors), the model performance was obviously enhanced by an increasing number of predictors. The most accurate estimates of the daily ETo for the study sites were achieved via the hybrid ANFIS-SFLA models using full predictors, with RMSE within 0.15 mm day−1, RRMSE within 4%, MAE within 0.11 mm day−1, and both a high R2 and NSE of 0.99 in the test phase at the two studied sites.

Download Full-text

Comparison of wavelet-based hybrid models for the estimation of daily reference evapotranspiration in different climates

Journal of Water and Climate Change ◽

10.2166/wcc.2018.113 ◽

2018 ◽

Vol 11 (1) ◽

pp. 39-53 ◽

Cited By ~ 1

Author(s):

Alireza Araghi ◽

Jan Adamowski ◽

Christopher J. Martinez

Keyword(s):

Good Accuracy ◽

Fuzzy Inference ◽

Reference Evapotranspiration ◽

Hybrid Models ◽

Detectable Effect ◽

Negative Effects ◽

Inference System ◽

Large Numbers ◽

Input Variables ◽

Different Climates

Abstract Reference evapotranspiration (ETo) is one of the most important factors in the hydrologic cycle and water balance studies. In this study, the performance of three simple and three wavelet hybrid models were compared to estimate ETo in three different climates in Iran, based on different combinations of input variables. It was found that the wavelet-artificial neural network was the best model, and multiple linear regression (MLR) was the worst model in most cases, although the performance of the models was related to the climate and the input variables used for modeling. Overall, it was found that all models had good accuracy in terms of estimating daily ETo. Also, it was found in this study that large numbers of decomposition levels via the wavelet transform had noticeable negative effects on the performance of the wavelet-based models, especially for the wavelet-adaptive network-based fuzzy inference system and wavelet-MLR, but in contrast, the type of db wavelet function did not have a detectable effect on the performance of the wavelet-based models.

Download Full-text

Establishing Coupled Models for Estimating Daily Dew Point Temperature Using Nature-Inspired Optimization Algorithms

Hydrology ◽

10.3390/hydrology9010009 ◽

2022 ◽

Vol 9 (1) ◽

pp. 9

Author(s):

Saeid Mehdizadeh ◽

Babak Mohammadi ◽

Farshad Ahmadi

Keyword(s):

Sunshine Duration ◽

Optimization Algorithms ◽

Hybrid Models ◽

Dew Point ◽

Estimation Accuracy ◽

Coupled Models ◽

Inference System ◽

Point Temperature ◽

Dragonfly Algorithm ◽

Dew Point Temperature

Potential of a classic adaptive neuro-fuzzy inference system (ANFIS) was evaluated in the current study for estimating the daily dew point temperature (Tdew). The study area consists of two stations located in Iran, namely the Rasht and Urmia. The daily Tdew time series of the studied stations were modeled through the other effective variables comprising minimum air temperature (Tmin), extraterrestrial radiation (Ra), vapor pressure deficit (VPD), sunshine duration (n), and relative humidity (RH). The correlation coefficients between the input and output parameters were utilized to determine the most effective inputs. Furthermore, novel hybrid models were proposed in this study in order to increase the estimation accuracy of Tdew. For this purpose, two optimization algorithms named bee colony optimization (BCO) and dragonfly algorithm (DFA) were coupled on the classic ANFIS. It was concluded that the hybrid models (i.e., ANFIS-BCO and ANFIS-DFA) demonstrated better performances compared to the classic ANFIS. The full-input pattern of the coupled models, specifically the ANFIS-DFA, was found to present the most accurate results for both the selected stations. Therefore, the developed hybrid models can be proposed as alternatives to the classic ANFIS to accurately estimate the daily Tdew.

Download Full-text

New Hybrids of ANFIS with Several Optimization Algorithms for Flood Susceptibility Modeling

Water ◽

10.3390/w10091210 ◽

2018 ◽

Vol 10 (9) ◽

pp. 1210 ◽

Cited By ~ 88

Author(s):

Dieu Tien Bui ◽

Khabat Khosravi ◽

Shaojun Li ◽

Himan Shahabi ◽

Mahdi Panahi ◽

...

Keyword(s):

Fuzzy Inference ◽

Model Performance ◽

Optimization Models ◽

Topographic Wetness Index ◽

Invasive Weed ◽

Inference System ◽

Conditioning Factors ◽

Flood Susceptibility ◽

Susceptibility Modeling ◽

Prediction Capability

This study presents three new hybrid artificial intelligence optimization models—namely, adaptive neuro-fuzzy inference system (ANFIS) with cultural (ANFIS-CA), bees (ANFIS-BA), and invasive weed optimization (ANFIS-IWO) algorithms—for flood susceptibility mapping (FSM) in the Haraz watershed, Iran. Ten continuous and categorical flood conditioning factors were chosen based on the 201 flood locations, including topographic wetness index (TWI), river density, stream power index (SPI), curvature, distance from river, lithology, elevation, ground slope, land use, and rainfall. The step-wise weight assessment ratio analysis (SWARA) model was adopted for the assessment of relationship between flood locations and conditioning factors. The ANFIS model, based on SWARA weights, was employed for providing FSMs with three optimization models to enhance the accuracy of prediction. To evaluate the model performance and prediction capability, root-mean-square error (RMSE) and receiver operating characteristic (ROC) curve (area under the ROC (AUROC)) were used. Results showed that ANFIS-IWO with lower RMSE (0.359) had a better performance, while ANFIS-BA with higher AUROC (94.4%) showed a better prediction capability, followed by ANFIS0-IWO (0.939) and ANFIS-CA (0.921). These models can be suggested for FSM in similar climatic and physiographic areas for developing measures to mitigate flood damages and to sustainably manage floodplains.

Download Full-text

Application of A Simple Landsat-MODIS Fusion Model to Estimate Evapotranspiration over A Heterogeneous Sparse Vegetation Region

Remote Sensing ◽

10.3390/rs11070741 ◽

2019 ◽

Vol 11 (7) ◽

pp. 741 ◽

Cited By ~ 11

Author(s):

Sajad Jamshidi ◽

Shahrokh Zand-Parsa ◽

Mojtaba Naghdyzadegan Jahromi ◽

Dev Niyogi

Keyword(s):

Vegetation Index ◽

Model Performance ◽

Growing Season ◽

Irrigation Scheduling ◽

Coefficient Of Determination ◽

In Situ Observation ◽

Successful Implementation ◽

Fusion Model ◽

Crop Phenology ◽

Sparse Vegetation

A simple Landsat-MODIS (Moderate Resolution Imaging Spectroradiometer) fusion model was used to generate 30-m resolution evapotranspiration (ET) maps for the 2010 growing season over a heterogeneous sparse vegetation, agricultural region using the METRIC (mapping evapotranspiration with internalized calibration) algorithm. The fusion model performance was evaluated, and experiments were undertaken to investigate the frequency for updating Landsat-MODIS data into the fusion model during the growing season, to maintain model accuracy and reduce computation. Initial evaluation of the fusion model resulted in high bias stemming from the landscape heterogeneity and small landholdings. To reduce the bias, the fusion model was modified to be applicable pixel-wise (i.e., implementing specific pixels for generating outputs), and an NDVI-based (Normalized Difference Vegetation Index) coefficient was added to capture crop phenology. A good agreement that resulted from the comparison of the fused and non-fused maps with root mean square error (RMSE) of 0.15 mm day−1 with coefficient of determination (R2) of 0.83 indicated successful implementation of the modifications. Additionally, the fusion model performance was evaluated against in-situ observation at the pixel level as well as the watershed level to estimate seasonal ET for the growing season. The default METRIC model (Landsat only) yielded relative error (RE) of 31% and RMSE of 2.44 mm day−1, while using the modified fusion model improved the accuracy resulting in RE of 3.5% with RMSE of 0.37 mm day−1. Considering different data frequency update, the optimal fusion experiment (RMSE of 0.61 mm day−1, and RE of 6.5%) required the consideration of the crop phenology and weekly updates in the early growing stage and harvest time, and bi-weekly for the rest of the season. The resulting fusion model for ET output is planned to be a part of ET mapping and irrigation scheduling systems.

Download Full-text

Results of the IACS Debris-covered Glaciers Working Group melt model intercomparison

10.5194/egusphere-egu2020-20639 ◽

2020 ◽

Author(s):

Francesca Pellicciotti ◽

Adria Fontrodona-Bach ◽

David Rounce ◽

Lindsey Nicholson

Keyword(s):

Energy Balance ◽

Working Group ◽

Model Performance ◽

Poor Performance ◽

Empirical Models ◽

Model Complexity ◽

Validation Data ◽

Energy Balance Models ◽

Model Components ◽

The Impact

Many mountain ranges across the globe support abundant debris-covered glaciers, and the proportion of glacierised area covered by debris is expected to increase under continuing negative mass balance. Within the activities of a newly established IACS Working Group (WG) on debris-covered glaciers, we have been carrying out an intercomparison of melt models for debris-covered ice, to identify the level of model complexity required to estimate sub-debris melt. This is a first necessary step to advance understanding of how debris impacts glacier response to climate at the local, regional, and global scale and accurately represent debris-covered glaciers in models of regional runoff and sea-level change projections.We compare ice melt rates simulated by 15 models of different complexity, forced at the point scale using data from nine automatic weather stations in distinct climatic regimes across the globe. We include energy-balance models with a variety of structural choices and model components as well as a range of simplified approaches. Empirical models are run twice: with values from literature and after recalibration at the sites. We then calculate uncertainty bounds for all simulations by prescribing a range of plausible parameters and varying them in a Monte Carlo framework. We restrict the comparison to the melt season and exclude conditions as few current models have the capability to account for them.Model results vary across sites considerably, with some sites where most models show a consistently good performance (e.g. in the Alps) which is also similar for energy-balance and empirical models, and sites where models diverge widely and the performance is overall poorer (e.g. in New Zealand and the Caucasus). It is also evident that with a few exceptions, most of the simpler, more empirical models have poor performance without recalibration. A few of the energy-balance models consistently give results different to the others, and we investigate structural differences, the impact of temporal resolution on the calculations (hourly versus daily) and the calculation of turbulent fluxes in particular. &#160;&#160;&#160;We provide a final assessment of model performance under different climate forcing, and evaluate models strengths and limitations against independent validation data from the same sites. We also provide suggestions for future model improvements and identify missing model components and crucial knowledge gaps and which require further attention by the debris-covered glacier community.

Download Full-text

A Wavelet Based Solar Radiation Prediction in Nigeria Using Adaptive Neuro-Fuzzy Approach

Indonesian Journal of Electrical Engineering and Computer Science ◽

10.11591/ijeecs.v12.i3.pp907-915 ◽

2018 ◽

Vol 12 (3) ◽

pp. 907 ◽

Cited By ~ 1

Author(s):

Sani Salisu ◽

Mohd Wazir Mustafa ◽

Mamunu Mustapha

Keyword(s):

Solar Radiation ◽

Meteorological Data ◽

Hybrid Approach ◽

Model Performance ◽

Coefficient Of Determination ◽

Maximum Temperature ◽

Percentage Error ◽

Anfis Model ◽

Inference System ◽

Neuro Fuzzy

In this study, a hybrid approach combining an Adaptive Neuro-Fuzzy Inference System (ANFIS) and Wavelet Transform (WT) is examined for solar radiation prediction in Nigeria. Meteorological data obtained from NIMET Nigeria comprising of monthly mean minimum temperature, maximum temperature, relative humidity and sunshine hours were used as inputs to the model and monthly mean solar radiation was used as the model output. The data used was divided into two for training and testing, with 70% used during the training phase and 30% during the testing phase. The hybrid model performance is assessed using three statistical evaluators, Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE) and Coefficient of determination (R2). According to the results obtained, a very accurate prediction was achieved by the WT- ANFIS model by improving the value of (R2) by at least 14% and RMSE by at least 78% when compared with other existing models. And a MAPE of 2% is recorded using the proposed approach. The obtained results prove the developed WT-ANFIS model as an efficient tool for solar radiation prediction.

Download Full-text

CPT to RVU conversion improves model performance in the prediction of surgical case length

Scientific Reports ◽

10.1038/s41598-021-93573-2 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Nicholas Garside ◽

Hamed Zaribafzadeh ◽

Ricardo Henao ◽

Royce Chung ◽

Daniel Buckland

Keyword(s):

Model Performance ◽

Individual Case ◽

Model Complexity ◽

Actual Case ◽

Surgical Case ◽

Relative Value ◽

Prediction Time ◽

Boosted Decision Tree ◽

Tree Models ◽

Duke University

AbstractMethods used to predict surgical case time often rely upon the current procedural terminology (CPT) code as a nominal variable to train machine-learned models, however this limits the ability of the model to incorporate new procedures and adds complexity as the number of unique procedures increases. The relative value unit (RVU, a consensus-derived billing indicator) can serve as a proxy for procedure workload and could replace the CPT code as a primary feature for models that predict surgical case length. Using 11,696 surgical cases from Duke University Health System electronic health records data, we compared boosted decision tree models that predict individual case length, changing the method by which the model coded procedure type; CPT, RVU, and CPT–RVU combined. Performance of each model was assessed by inference time, MAE, and RMSE compared to the actual case length on a test set. Models were compared to each other and to the manual scheduler method that currently exists. RMSE for the RVU model (60.8 min) was similar to the CPT model (61.9 min), both of which were lower than scheduler (90.2 min). 65.2% of our RVU model’s predictions (compared to 43.2% from the current human scheduler method) fell within 20% of actual case time. Using RVUs reduced model prediction time by ninefold and reduced the number of training features from 485 to 44. Replacing pre-operative CPT codes with RVUs maintains model performance while decreasing overall model complexity in the prediction of surgical case length.

Download Full-text

Global Analysis of Atmospheric Transmissivity Using Cloud Cover, Aridity and Flux Network Datasets

Remote Sensing ◽

10.3390/rs13091716 ◽

2021 ◽

Vol 13 (9) ◽

pp. 1716

Author(s):

Ankur Srivastava ◽

Jose F. Rodriguez ◽

Patricia M. Saco ◽

Nikul Kumari ◽

Omer Yetemen

Keyword(s):

Cloud Cover ◽

Climatic Factors ◽

Global Analysis ◽

Empirical Relationship ◽

Aridity Index ◽

Global Scale ◽

Empirical Models ◽

Coefficient Of Determination ◽

Crop Modeling ◽

Atmospheric Transmissivity

Atmospheric transmissivity (τ) is a critical factor in climatology, which affects surface energy balance, measured at a limited number of meteorological stations worldwide. With the limited availability of meteorological datasets in remote areas across different climatic regions, estimation of τ is becoming a challenging task for adequate hydrological, climatic, and crop modeling studies. The availability of solar radiation data is comparatively less accessible on a global scale than the temperature and precipitation datasets, which makes it necessary to develop methods to estimate τ. Most of the previous studies provided region specific datasets of τ, which usually provide local assessments. Hence, there is a necessity to give the empirical models for τ estimation on a global scale that can be easily assessed. This study presents the analysis of the τ relationship with varying geographic features and climatic factors like latitude, aridity index, cloud cover, precipitation, temperature, diurnal temperature range, and elevation. In addition to these factors, the applicability of these relationships was evaluated for different climate types. Thus, empirical models have been proposed for each climate type to estimate τ by using the most effective factors such as cloud cover and aridity index. The cloud cover is an important yet often overlooked factor that can be used to determine the global atmospheric transmissivity. The empirical relationship and statistical indicator provided the best performance in equatorial climates as the coefficient of determination (r2) was 0.88 relatively higher than the warm temperate (r2 = 0.74) and arid regions (r2 = 0.46). According to the results, it is believed that the analysis presented in this work is applicable for estimating the τ in different ecosystems across the globe.

Download Full-text

Modelling response of infiltration loss toward water table depth using RBFN, RNN, ANFIS techniques

International Journal of Knowledge-based and Intelligent Engineering Systems ◽

10.3233/kes-210066 ◽

2021 ◽

Vol 25 (2) ◽

pp. 227-234

Author(s):

Sandeep Samantaray ◽

Abinash Sahoo

Keyword(s):

Neural Network ◽

Mean Square Error ◽

Water Table ◽

Environmental Sustainability ◽

Model Development ◽

Water Table Depth ◽

Water Levels ◽

Coefficient Of Determination ◽

Mean Square ◽

Inference System

Accurate prediction of water table depth over long-term in arid agricultural areas are very much important for maintaining environmental sustainability. Because of intricate and diverse hydrogeological features, boundary conditions, and human activities researchers face enormous difficulties for predicting water table depth. A virtual study on forecast of water table depth using various neural networks is employed in this paper. Hybrid neural network approach like Adaptive Neuro Fuzzy Inference System (ANFIS), Recurrent Neural Network (RNN), Radial Basis Function Neural Network (RBFN) is employed here to appraisal water levels as a function of average temperature, precipitation, humidity, evapotranspiration and infiltration loss data. Coefficient of determination (R2), Root mean square error (RMSE), and Mean square error (MSE) are used to evaluate performance of model development. While ANFIS algorithm is used, Gbell function gives best value of performance for model development. Whole outcomes establish that, ANFIS accomplishes finest as related to RNN and RBFN for predicting water table depth in watershed.

Download Full-text

Development of a semi-parametric PAR (Photosynthetically Active Radiation) partitioning model for the United States, version 1.0

Geoscientific Model Development ◽

10.5194/gmd-7-2477-2014 ◽

2014 ◽

Vol 7 (5) ◽

pp. 2477-2484 ◽

Cited By ~ 3

Author(s):

J. C. Kathilankal ◽

T. L. O'Halloran ◽

A. Schmidt ◽

C. V. Hanson ◽

B. E. Law

Keyword(s):

United States ◽

Elevation Angle ◽

Model Performance ◽

Diffuse Radiation ◽

The United States ◽

Polynomial Model ◽

Coefficient Of Determination ◽

Data Set ◽

Cubic Polynomial ◽

Solar Elevation Angle

Abstract. A semi-parametric PAR diffuse radiation model was developed using commonly measured climatic variables from 108 site-years of data from 17 AmeriFlux sites. The model has a logistic form and improves upon previous efforts using a larger data set and physically viable climate variables as predictors, including relative humidity, clearness index, surface albedo and solar elevation angle. Model performance was evaluated by comparison with a simple cubic polynomial model developed for the PAR spectral range. The logistic model outperformed the polynomial model with an improved coefficient of determination and slope relative to measured data (logistic: R2 = 0.76; slope = 0.76; cubic: R2 = 0.73; slope = 0.72), making this the most robust PAR-partitioning model for the United States currently available.

Download Full-text