Willmott Index Of Agreement Research Articles

Calibration and validation of the AquaCrop model for forage cactus production systems under different management interventions in the semi-arid region of Brazil

Simulation models are useful tools for estimating plant response and contributing to the development of management strategies and yield predictions in crops. This study aimed to calibrate and evaluate the AquaCrop-FAO model for the forage cactus imposed on different cropping systems in the semi-arid region of Brazil. Four experiments were conducted: cactus-sorghum intercropping system with five spacings between plants (0.10, 0.20, 0.30, 0.40 and 0.50 m); cactus-sorghum intercropping system with five spacings between rows (1.00, 1.25, 1.50 and 1.75 m); single cactus with four levels of mulch (0, 5, 10 and 15 Mg ha−1); and single forage cactus with four levels of nitrogen fertilisation (50, 150, 300 and 450 kg of N ha−1). The performance of the model was evaluated using the following parameters: root mean squared error (RMSE); normalised RMSE (NRMSE), coefficients of determination (R2), Nash-Sutcliffe model efficiency coefficient (ME) and the Willmott index of agreement (d). During the calibration stage, the statistical indices pointed to the good performance of the AquaCrop model in estimating dry biomass in most of the cropping systems, with 10 < NRMSE <20; R2 > 0.96 and ME > 0.95. The AquaCrop model showed the best performance simulating systems under different levels of nitrogen fertilisation. Among the intercropping systems, the denser arrangements showed the highest values for simulated biomass.

Leaf area prediction of sweet potato cultivars: An approach to a non-destructive and accurate method

Many researchers consider the leaf area the most critical parameter for plant growth and physiology studies. Using allometric equations is considered an effective and non-destructive method for estimating the leaf area of plants. The objectives of this study were to construct allometric equations to predict the leaf area of sweet potato cultivars as a function of leaf dimensions and to verify the effectiveness of the allometric equations constructed for each cultivar individually and for the grouped data. Two hundred fifty leaves of the sweet potato cultivars Beauregard, BRS Amélia, BRS Cuia, and Coquinho were collected, and then the length and width of each leaf were measured. Subsequently, the product between leaf length and width and the actual leaf area was calculated. Linear, linear without intercept, power, and exponential regression models were used to construct allometric equations to predict leaf area. The best equations were chosen according to the highest coefficients of determination and correlation, Willmott agreement index, and lowest Akaike information criterion and root mean square error. The equations obtained with the linear models without intercept and power using the length and width of the leaves were the most indicated for predicting the leaf area of sweet potato cultivars.

Allometric modelling to estimate the leaf area of purslane (Portulaca umbraticola Kunth.)

Portulaca umbraticola is a species that has several uses, including ornamental. Among the variables measured in ornamental plants are the biometric data of the leaves, which can be obtained through image analysis, using indirect method, which is simple, accurate and economically viable. This work aimed to propose equations obtained by mathematical models to estimate the leaf area of P. umbraticola. The study was conducted at the Universidade Federal Rural do Semi-Árido, Mossoró, RN, Brazil. Three hundred leaves of each of the five genotypes of P. umbraticola were evaluated for length, width, product between length and width, and leaf area. Determination and correlation coefficients, Akaike information criterion, Willmott agreement index, coefficient of variation, normality using the Shapiro-Wilk test, principal components and Student's t-test were used to systematize the data. Descriptive measurements obtained from the leaf blades of P. umbraticola showed variability between genotypes, demonstrating that the sample was representative for the population analyzed. The regression models obtained were linear and potential. Allometric equations can be used to estimate leaf area in P. umbraticola, and the recommended equations are ŷ = 0.356*L (R² = 0.91) for length, ŷ = 1.665*W (R² = 0.90) for width and ŷ = 0.016 × 0.0711*LW (R² = 0.99), ŷ = 0.715*LW (R² = 0.99) and ŷ = 0.724*LW (R² = 0.99) for the relationship between leaf length and width. The equations to estimate the leaf area of P. umbraticola showed excellent data fits, with values of R² = 0.993 and little dispersion of the data.

Groundwater salinization risk assessment using combined artificial intelligence models.

Assessing the risk of groundwater contamination is of crucial importance for the management of water resources, particularly in arid regions such as Menzel Habib (south-eastern Tunisia). The aim of this research is to create and validate artificial intelligence models based on the original DRASTIC vulnerability methodology to explain groundwater salinization risk (GSR). To this end, several algorithms, such as artificial neural networks (ANN), support vector regression (SVR), and multiple linear regression (MLR), were applied to the Menzel Habib aquifer system. The results obtained indicate that the DRASTIC Vulnerability Index (VI) ranges from 91 to 141 and is classified into two categories: low and moderate vulnerability. However, the correlation between groundwater total dissolved solids (TDS) and the Vulnerability Index is relatively weak (r < 0.5). Indeed, the original DRASTIC index needs some improvements. To improve it, some adjustments are required, notably by incorporating the TDS-groundwater salinization risk (GSR) indicator. The seven parameters of the original DRASTIC model were used as inputs for the artificial intelligence models, while the GSR values were used as outputs. Performance indicators, such as the correlation coefficient (r) and the Willmott Agreement Index (d), showed that the ANN model outperformed the SVR and MLR models. Indeed, during the training phase, the ANN model obtained r values equal to 0.89 and d values of 0.4, demonstrating the superiority, robustness, and accuracy of ANN-based methodologies over the original DRASTIC model. The findings could provide valuable information to guide management of groundwater contamination risks, especially in arid regions.

Comparative study of different machine learning approaches for predicting the compressive strength of palm fuel ash concrete

Nowadays, the challenge of industrialising concrete is to produce economical, ecological and efficient concrete. As a result, cementitious additions are increasingly used in concrete binders with proportions conferring the desired properties, particularly compressive strength. Palm oil fuel ash (POFA) from palm trees is one of these additions and introduced by substitution of cement in concrete. Although laboratory tests applied to concrete samples are expensive and time-consuming, considerable research is directed towards modelling approaches and machine learning (ML) applications. The purpose of this study is to explore different ML techniques, including artificial neural network (ANN), ANN with combined inputs (ANNX), particle swarm optimisation (PSO) and genetic algorithm (GA). These techniques are used to predict the compressive strength of POFA concrete derived from a dataset consisting of 249 samples with regard to 6 input parameters, namely, cement and POFA contents, aggregate ratio, water-to-binder ratio, superplasticiser dosage and curing duration. The performance of these techniques is evaluated using six performance indices: correlation coefficient, Willmott index of agreement, Nash–Sutcliffe efficiency, root mean squared error, mean absolute error and mean squared relative error. The comparison shows that ANN and ANNX have better predictive accuracy than PSO and GA, and ANNX performs relatively better than ANN. Results are presented using a 3D surface response plot which visually depicts the relationship between predicted outcomes and influential parameters. The comparison made between ANNX and other models in the literature demonstrated that the present model is reliable to predict CS of POFA concrete with high precision in a wide range of data.

Modelos alométricos para estimativa não destrutiva da área de folíolos da umbu-cajazeira (Spondias sp.)

ABSTRACT The umbu-cajazeira (Spondias sp.) is a fruit tree native to the semi-arid of Brazil, with significant environmental and agro-socioeconomic potential for the region. Determining leaflet area (LA) is essential for understanding the interactions between plant growth, physiology, nutrition, and environment. Thus, this research aims to construct allometric equations that non-destructively estimate the umbu-cajazeira (Spondias sp.) leaflet area, considering the leaflet dimensions (length and width). The leaflets were collected from mother plants of umbu-cajazeira (Spondias sp.) and individually measured for length (L), width (W), and LA using digitized images. These data were submitted for descriptive and regression analysis. The LA was estimated using, linear power, and exponential regression models based on the L and W of the leaflets. The best model and equation were chosen based on the following selection criteria: coefficient of determination, Willmott agreement index, root mean square error, and mean absolute error. The power model using LW presented the best equation ŷ = 0.76 * LW0.98 to estimate the leaflet area of umbu-cajazeira (Spondias sp.). This study provides a reliable, accurate, fast, and non-destructive approach for agronomic researchers and growers to determine the LA of the species.

Hybrid model combining LSTM with discrete wavelet transformation to predict surface methane concentration in the Arctic Island Belyy

In this paper, we propose a hybrid model that combines wavelet transformation of the raw data and an artificial neural network with long short-term memory (LSTM). This model allows researchers to increase the accuracy of time series forecasting.The model is based on data from environmental monitoring of greenhouse gases on the Belyy Island of the Yamal-Nenets Autonomous Okrug, Russia.The raw data for building the proposed model were obtained in July–August 2016. The CH4 concentration time series was decomposed using a discrete wavelet transform into four components - one approximating and three detailing. These components, along with a timed temperature series, were used to train six ANNs—two exogenous input autoregressive networks (NAR) and four LSTM networks. The forecast was calculated as the sum of forecasts for each of the components. Forecast accuracy was assessed using several indices (mean absolute error (MAE), root mean square error (RMSE), mean square relative error (RMSRE), Willmott's agreement index (IA1, IA2)) and a Taylor diagram. The hybrid model based on LSTM showed the best accuracy. The hybrid model based on LSTM showed the best accuracy. MAE, RMSE, and RMSRE errors decreased by more than 70%. In terms of IA1 and IA2, the models improved by 11% and 30% (when comparing the best hybrid models LSTM and NAR. For the hybrid LSTM model errors RMSE, RMSRE, and MAE are more than 20% less than for the base LSTM model for training in which data without wavelet transform were used. Willmott agreement indices (IA1, IA2) rose from 32% to 59%, depending on the models being compared. So, the increase in accuracy after applying the described approach was up to 79% for models based on LSTM (more precisely, from 20% to 79%, depending on the indicator). Applying wavelet transform data to train the NAR-based model reduced MAE, RMSE, and RMSRE errors by 27%, 30%, and 31%, respectively. The Willmott agreement indices (IA1, IA2) rose from 41% to 45% respectively. The Taylor diagram also shows the advantage of the proposed approach.

Determination of Soil Moisture for Sandy Soils Using Gas and Microwave Ovens

Since there are several methods used to determine soil moisture, the objective was to compare the moisture values obtained with the use of a gas oven and a microwave to the values obtained with the standard oven method. Having been carried out at the Hydraulics Laboratory of the State University of Maringá, Umuarama Regional Campus, the performance of the methods was evaluated in the determination of five levels of moisture (10, 15, 20, 25, and 30%), with dry masses obtained at 15, 20, 25 and 30 minutes for the gas oven and at 2, 4, 6 and 8 minutes for the microwave. The best drying times were selected using the Taylor Diagram. Linear regression was performed between the methods, in order to obtain the correlation coefficient (r), the Willmott agreement index (d), the performance index (c), absolute error (EA), absolute mean error (ERA), and root mean square error (RQME). By analyzing the results obtained, it can be stated that the gas oven method and the microwave method presented excellent performance when compared to the standard method. It was also concluded that the time of 25 minutes for the gas oven method was efficient and, for the microwave method, the drying time of 8 minutes was sufficient, meaning a reduction in the time for moisture determination.

A non-destructive method for predicting the leaflet area of Cassia fistula L.: An approach to regression models

The leaf area is an important parameter for plant growth and development, usually determined by direct and destructive methods. In these methods, high-cost equipment is mainly used, requiring more time and restrictions in the analyses. Therefore, the study sought to construct, through regression models, allometric equations that allow estimating the area of leaflets of Cassia fistula L. by a non-destructive method using the linear dimensions of the leaflets. Six hundred leaflets of different sizes and shapes were randomly collected in adult species matrices. Each leaflet's length, width, and leaf area were measured using digital images. Next, the products were calculated between length and length, width and width, and length and width. Two regression models were used to construct five allometric equations to predict the area of leaflets of C. fistula. All the proposed equations can be used to estimate the leaf area of the species; however, the best equations were constructed using the power and linear without intercept models, being the most indicated to estimate the leaf area of the species, presenting the highest determination coefficients (R² = 0.9949 and 0.9993), Pearson's correlation coefficients (r = 0.9374 and 0.9374), Willmott's agreement index (d = 0.9987 and 0.9986), and lowers Akaike information criterion (AIC = 1071.3 and 1098.2) and normalized root mean square error (NRMSE = 7.2 and 7.1). Thus, the models y^=0.747*LW0.981 and y^=0.696*LW should be applied to estimate the area of leaflets of C. fistula.

Forecasting of stage-discharge in a non-perennial river using machine learning with gamma test

Knowledge of the stage-discharge rating curve is useful in designing and planning flood warnings; thus, developing a reliable stage-discharge rating curve is a fundamental and crucial component of water resource system engineering. Since the continuous measurement is often impossible, the stage-discharge relationship is generally used in natural streams to estimate discharge. This paper aims to optimize the rating curve using a generalized reduced gradient (GRG) solver and the test the accuracy and applicability of the hybridized linear regression (LR) with other machine learning techniques, namely, linear regression-random subspace (LR-RSS), linear regression-reduced error pruning tree (LR-REPTree), linear regression-support vector machine (LR-SVM) and linear regression-M5 pruned (LR-M5P) models. An application of these hybrid models was performed and test to modeling the Gaula Barrage stage-discharge problem. For this, 12-year historical stage-discharge data were collected and analyzed. The 12-year historical daily flow data (m3/s) and stage (m) from during the monsoon season, i.e., June to October only from 03/06/2007 to 31/10/2018, were used for discharge simulation. The best suitable combination of input variables for LR, LR-RSS, LR-REPTree, LR-SVM, and LR-M5P models was identified and decided using the gamma test. GRG-based rating curve equations were found to be as effective and more accurate as conventional rating curve equations. The outcomes from GRG, LR, LR-RSS, LR-REPTree, LR-SVM, and LR-M5P models were compared to observed values of daily discharge based on Nash Sutcliffe model efficiency coefficient (NSE), Willmott Index of Agreement (d), Kling-Gupta efficiency (KGE), mean absolute error (MAE), mean bias error (MBE), relative bias in percent (RE), root mean square error (RMSE) Pearson correlation coefficient (PCC) and coefficient of determination (R2). The LR-REPTree model (combination 1: NSE = 0.993, d = 0.998, KGE = 0.987, PCC(r) = 0.997, and R2 = 0.994 and minimum value of RMSE = 0.109, MAE = 0.041, MBE = −0.010 and RE = −0.1%; combination 2; NSE = 0.941, d = 0.984, KGE = 0. 923, PCC(r) = 0. 973, and R2 = 0. 947 and minimum value of RMSE = 0. 331, MAE = 0.143, MBE = −0.089 and RE = −0.9%) performed superior to the GRG, LR, LR-RSS, LR-SVM, and LR-M5P models in all input combinations during the testing period. It was also noticed that the performance of the alone LR and its hybrid models (i.e., LR-RSS, LR-REPTree, LR-SVM, and LR-M5P) was better than the conventional stage-discharge rating curve, including the GRG method.

AquaCrop Model Performance in Yield, Biomass, and Water Requirement Simulations of Common Bean Grown under Different Irrigation Treatments and Sowing Periods

The application of crop growth simulation and water management models will become increasingly important in the future. They can be used to predict yield reductions due to water scarcity and allocate water to ensure profitable crop production. The objective of this research was to calibrate the AquaCrop model for common bean (Faseolus vulgaris L.) grown in temperate climates and to test whether the model can be used for different irrigation strategies to achieve high yield productivity. The model was calibrated using data obtained from two years of experimental research in the Serbian territory of the Syrmia region. There were three sowing periods/plots: I—mid April, II—end of May/beginning of June, and III—third decade of June/beginning of July; and three levels of irrigation/subplots: full irrigation (F) providing 100% of crop evapotranspiration (ETc), mild deficit irrigation (R) at 80% of ETc, and moderate deficit irrigation (S) at 60% of ETc. The results show that the AquaCrop model accurately predicts common bean yield, biomass, canopy cover, and water requirements. The statistical indices of the calibrated dataset, coefficient of determination (R2), normalized root mean square error (NRMSE), mean bias error (MBE), and Willmott agreement index (d) for yield and biomass were: 0.91, 0.99; 6.9%, 11.4%; −0.046, 1.186 and 0.9, 0.89, respectively. When testing three irrigation strategies, the model accurately predicted irrigation requirements for the full and two deficit irrigation strategies, with only 29 mm, 32 mm, and 34 mm more water than was applied for the Fs, Rs, and Ss irrigation strategy, respectively. The AquaCrop model performed well in predicting irrigated yield and can be used to estimate the yield of common bean for different sowing periods and irrigation strategies.

Modeling retroreflectivity degradation of traffic signs using artificial neural networks

Traffic signs are vital for communicating guidance, rules, warning, and other highway agency information for safe and efficient navigation through transportation networks. Signs must be clearly detectable, readable, and understandable to fulfill their intended purpose. Poor sign visibility, particularly during nighttime, is the leading cause of fatalities worldwide. Sign retroreflectivity is one of the key measures to evaluate sign visibility conditions. It gradually deteriorates over time with sign aging, exposure, and other environmental conditions necessitating periodic sign maintenance or, ultimately, replacement when the sign retro values fall below the minimum prescribed standards. In literature, studies have mostly used traditional statistical regression models to model sign retroreflectivity as a function of available explanatory variables. Further, these studies have proposed separate retro degradation prediction models for different sign sheeting grades and colors that limit their applications for other scenarios. To fill the research gap, this study compared the performance of the linear regression method with three different architecture of the neural network namely, Feed-Forward Neural Network (FFNN), Cascade Forward Neuran Network (CFNN), and Elman neural networkNeural Network (ELMNN) for signs retro prediction with an aim to optimize sign maintenance and replacement activities and to enhance road safety. All the Neural Network models were employed with varying combinations of training algorithms, activation functions, and model parameters. Sign retro data for 539 in-service signs along selected sections of two expressways (M-1 and M-2) near the capital city of Islamabad in Pakistan were collected through portable handheld retroreflectometer GR3. Data on other sign attributes like sign ages (0, 2, 5, and 10 years), sign orientation, observation angle, sign sheeting brand, grade, and color were also acquired. Feature-based sensitivity analysis was conducted to identify the relative importance and ranking of input predictor variables. Model prediction results expressed in terms of various statistical evaluation metrics root mean square error (RMSE), mean absolute percent error (MAPE), RMSE-observation standard deviation satio (RSR), coefficient of determination (R2), Willmott's index of agreement (WIA), Nash-Sutcliffe efficiency (NSEC), and percent bias (PBIAS) showed that all the NN models outperformed the regression technique. Comparing the NN models, ELMNN architecture with 21 neurons in the hidden layer for ‘tansig’ activation function and ‘trainlm’ training algorithm yielded better retro-prediction performance. Feature sensitivity analysis revealed that variables sign age, sheeting brand, color, and observation angle were the most sensitive variables in predicting the retro output. Findings of this study can guide the transport agencies and decision-makers for effective policy implications and sign management practices.

Efficacy of linear multiple regression and artificial neural network for long-term rainfall forecasting in Western Australia

Precipitation is one of the most intrinsic resources for manifold industrial activities all over Western Australia; consequently, immaculate rainfall prediction is indispensable for flood mitigation as well as water resources management. This study investigated the performance of artificial neural networks (ANN) and Linear multiple regression (LMR) analysis to forecast long-term seasonal spring rainfall in Western Australia, using lagged El Nino Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) as potential climatic phenomena. The ANN was developed in the form of multilayer perceptron using Levenberg–Marquardt algorithm and subsequently LMR was used with statistical significance for future spring rainfall forecast. The total climatic dataset has been divided into calibration and testing phases to determine the efficacy of the developed models. Different statistical skill tests such as root mean square error (RMSE), mean absolute error (MAE), and Willmott index of agreement ‘d’ were used to assess the efficacy of LMR and ANN modelling. In general, LMR has lower MAE and RMSE values as compared to ANN for most of the stations during calibration and testing periods, whereas ANN models performed better than LMR models based on ‘d’ values. The overall statistical analysis paradigm suggests the efficacy of LMR over ANN models for rainfall forecasting using more climatic variables. As a result, the developed LMR model, incorporated with lagged global climate indices, will facilitate the adequate preparedness for the risks associated with potential droughts in the study region.

Adapting the CROPGRO‐faba bean model to simulate the growth and development of Amaranthus species

AbstractThe aim of this study was to adapt the CROPGRO model to simulate growth and development processes of Amaranthus spp. under central European conditions. In 2017 and 2018, two field experiments with two amaranth cultivars (grain type, A. hypochondriacus L. Neuer Typ [NT]; fodder type, A. caudatus L. K63 [K63]) were conducted in southern Germany. Based on experimental and literature data, parameter coefficients that drive physiological processes at species, cultivar, and ecotype levels were calibrated to predict the time series experimental observations of various growth and development traits. Statistical evaluation of the model adaptation was performed using root mean square error (RMSE, in variable units, 0 equals perfect fit) and the Willmott agreement index (d‐Stat., range from 0 to 1, 1 equals perfect fit). For NT and K63, respectively, the model adaptation led to accurate predictions of canopy height (RMSE, 0.07 and 0.24 m; d‐Stat.,0.98 and 0.92), panicle weight (RMSE, 2,034 and 1,153 kg ha–1; d‐Stat., 0.92 and 0.94), panicle harvest index (RMSE, 0.05 and 0.06; d‐Stat., 0.99 and 0.96), leaf N concentration (RMSE, 0.38 and 0.40%; d‐Stat., 0.94 and 0.92) and aboveground biomass (RMSE, 2,948 and 2,572 kg ha–1; d‐Stat., 0.88 and 0.91). In summary, the CROPGRO model was successfully adapted for Amaranthus spp. The adapted model can be further improved as it is made available for evaluation in different locations and environments including limited soil N supply.

Predicting daily soil temperature at multiple depths using hybrid machine learning models for a semi-arid region in Punjab, India.

Prediction of soil temperature (ST) at multiple depths is important for maintaining the physical, chemical, and biological activities in soil for various scientific aspects. The present study was conducted in a semi-arid region of Punjab to predict the daily ST at 5-cm (ST5), 15-cm (ST15), and 30-cm (ST30) soil depths by employing the three-hybrid machine learning (ML) paradigms, i.e. support vector machine (SVM), multilayer perceptron (MLP), adaptive neuro-fuzzy inference system (ANFIS) optimized with slime mould algorithm (SMA), particle swarm optimization (PSO), and spotted hyena optimizer (SHO) algorithms. Five scenarios with different input variables were constructed using daily meteorological parameters, and the optimal one was extracted by exploiting the GT (gamma test). The feasibility of the proposed hybrid SVM, MLP, and ANFIS models was inspected based on performance metrics and visual interpretation. According to the results, the SVM-SMA model yields better estimates than other models at 5-cm, 15-cm, and 30-cm soil depths, respectively, for scenario 5 in the validation phase. Furthermore, conferring to the results, the SMA algorithm-based SVM model had lower (higher) values of mean absolute error, root mean square error, and index of scattering (Nash-Sutcliffe efficiency, coefficient of correlation, and Willmott index of agreement) and proved the better feasibility of SVM models in predicting daily ST at multiple depths on the study site.

Estimating leaf area of basil cultivars through linear dimensions of leaves

Ocimum basilicum L. (basil) is an annual herb belonging to the Lamiaceae family that has economic importance for many regions around the world. Thus, ecophysiological studies are needed to assess this species growth and dispersal. This work aimed to obtain equations from regression models that meaningfully estimate the leaf area of ​​basil cultivars using linear dimensions of leaves. For this purpose, 300 leaves from 'Italiano Roxo' and 500 leaves from 'Folha Fina' cultivar were collected from plants cultivated in a greenhouse. Then, the length, width, and leaf area of each leaf were measured, and product of length by width were calculated. The equations were adjusted using the simple linear, linear without intercept, quadratic, cubic, power, and exponential regression models. Criteria for selecting the best equations were highest determination coefficient and Willmott's agreement index, lowest Akaike information criterion and root mean square error, and BIAS index closest to zero. All the equations fitted using the product of length by width (L.W) can estimate the leaf area of basil cultivars. Thus, basil leaf area can be estimated through a non-destructive method using linear dimensions of leaves. However, the equation ŷ = 0.8175*LW0.9307 is the most suitable for 'Italiano Roxo' and ŷ = 0.6335*LW for 'Folha Fina'.

Estimation of Thunbergia grandiflora leaf area from allometric models

Sky vine (Thunbergia grandiflora Roxb) is a vine with important structural components for forest environments. Studies on growth and development are necessary, because of the environmental and economic importance. The leaf area determination is essential for ecophysiological studies to understand the relationship of the plant with the environment. The objective of this work was to estimate an allometric equation to estimate the leaf area of T. grandiflora from linear dimensions. 200 leaves of different shapes and sizes were collected from adult plants and the length (L), width (W), the product between length and width (LW), and real leaf area (LA) were measured. The linear regression, linear without intercept, quadratic, cubic, power, and exponential models were used to estimate the equations. The criteria for determining the best model were higher determination coefficient (R2), Willmott's agreement index (d), lower Akaike information criterion (AIC), the root of the mean error square (RMSE), and BIAS index closer to zero. The leaf area of T. grandiflora can be estimated satisfactorily by the equation ŷ = 0.58*LW.

Modeling Soil Temperature for Different Days Using Novel Quadruplet Loss-Guided LSTM.

Soil temperature (Ts), a key variable in geosciences study, has generated growing interest among researchers. There are many factors affecting the spatiotemporal variation of Ts, which poses immense challenges for the Ts estimation. To enrich processing information on loss function and achieve better performance in estimation, the paper designed a new long short-term memory model using quadruplet loss function as an intelligence tool for data processing (QL-LSTM). The model in this paper combined the traditional squared-error loss function with distance metric learning between the sample features. It can zoom analyze the samples accurately to optimize the estimation accuracy. We applied the meteorological data from Laegern and Fluehli stations at 5, 10, and 15 cm depth on the 1st, 5th, and 15th day separately to verify the performance of the proposed soil temperature estimation model. Meanwhile, this paper inputs the variables into the proposed model including radiation, air temperature, vapor pressure deficit, wind speed, air pressure, and past Ts data. The performance of the model was tested by several error evaluation indices, including root mean square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe model efficiency coefficient (NS), Willmott Index of Agreement (WI), and Legates and McCabe index (LMI). As the test results at different soil depths show, our model generally outperformed the four existing advanced estimation models, namely, backpropagation neural networks, extreme learning machines, support vector regression, and LSTM. Furthermore, as experiments show, the proposed model achieved the best performance at the 15 cm depth of soil on the 1st day at Laegern station, which achieved higher WI (0.998), NS (0.995), and LMI (0.938) values, and got lower RMSE (0.312) and MAE (0.239) values. Consequently, the QL-LSTM model is recommended to estimate daily Ts profiles estimation on the 1st, 5th, and 15th days.

Allometric equations to estimate the leaf area of Psychotria colorata (Willd. Ex Schult.) Müll.Arg.

Estimating leaf area using non-destructive methods from regression equations has become a more efficient, quick, and accurate way. Thus, this study aimed to propose an equation that significantly estimates the leaf area of Psychotria colorata (Rubiaceae) through linear leaf dimensions. For this purpose, 200 leaves of different shapes were collected, and length (L), width (W), product of length by width (L.W), and real leaf area (LA) of each leaf blade were determined. Then, equations were adjusted for predicting leaf area using simple linear, linear (0.0), quadratic, cubic, power, and exponential regression models. The proposed equation was selected according to the coefficient of determination (R²), Willmott's agreement index (d), Akaike's information criterion (AIC), mean absolute error (MAE), mean squared error (RMSE) and BIAS index. It was noted that the equations adjusted using L.W met the best criteria for estimating leaf area, but the equation LA = 0.59 * L.W from linear regression without intercept was the most suitable. This equation predicts that 59% of leaf area is explained by L.W. Concluding, the leaf area of P. colorata can be estimated using an allometric equation that uses linear leaf blade dimensions.

Can intercropping be an adaptation to drought? A model‐based analysis for pearl millet–cowpea

AbstractCereal–legume intercropping is promoted within semi‐arid regions as an adaptation strategy to water scarcity and drought for low‐input systems. Our objectives were firstly to evaluate the crop model APSIM for pearl millet (Pennisetum glaucum (L.))—cowpea (Vigna unguiculata (L.) Walp) intercropping—and secondly to investigate the hypothesis that intercropping provides complimentary yield under drought conditions. The APSIM model was evaluated against data from a two year on station field experiment during the dry season of a semi‐arid environment in Patancheru, India, with severe, partial and no water deficit stress (well‐watered); densities of 17 and 33 plants per m−2, and intercrop and sole crop production of pearl millet and cowpea. Overall, APSIM captured the dynamics of grain yields, indicated by the Willmott Index of Agreement (IA: 1 optimal, 0 the worst) 0.91 from 36 data points (n), total biomass (IA: 0.90, n = 144), leaf area index (LAI, IA = 0.77, n = 66), plant height (IA 0.96, n = 104 pearl millet) and cowpea (IA 0.81, n = 102), as well as soil water (IA 0.73, n = 126). Model accuracy was reasonable in absolute terms (RMSE pearl millet 469 kg/ha and cowpea 322 kg/ha). However, due to low observed values (observed mean yield pearl millet 1,280 kg/ha and cowpea 555 kg/ha), the relative error was high, a known aspect for simulation accuracy in low‐yielding environments. The simulation experiment compared the effect of intercropping pearl millet and cowpea versus sole cropping under different plant densities and water supplies. A key finding was that intercropping pearl millet and cowpea resulted in similar total yields to the sole pearl millet. Both sole and intercrop systems responded strongly to increasing water supply, except sole cropped cowpea, which performed relatively better under low water supply. High plant density had a consistent effect, leading to lower yields under low water supply. Higher yields were achieved under high density, but only when water supply was high: absolute highest total intercrop yields were 4,000 (high density) and 3,500 kg/ha (low density). This confirms the suitability of the common practice among farmers who use the low planting density under water scarce conditions. Overall, this study confirms that intercropping is no silver bullet, i.e. not per se a way to achieve high yield production or reduce risk under drought. It does, however, provide an opportunity to diversify food production by additionally integrating protein rich crops, such as cowpea.