Mean Bias Research Articles

Measurement of 1,25-dihydroxyvitamin D in serum by LC-MS/MS compared to immunoassay reveals inconsistent agreement in paediatric samples.

Automated immunoassays for 1,25-dihydroxyvitamin D (1,25(OH)2D) have increased the use of serum measurements in clinical and research settings, but disagreement with LC-MS/MS methods remains an issue. In this study, we examined this problem using samples obtained from healthy young adults, n=80, mean age 21.7 (18-32) years, and a large cohort of paediatric samples, n=422, mean age 7.3 (0-17) years. We compared serum concentrations of 1,25(OH)2D3/D2 produced by the DiaSorin LIAISON® XL immunoassay against an LC-MS/MS method with immunoaffinity enrichment and DAPTAD derivation. Both assays showed intra/inter-assay imprecision of≤9.4 % across their respective assay range. DEQAS between April 2020 to Jan 2024 (n=80) showed mean bias (SD, 95 %CI) for DiaSorin-0.6 % (6.2,-12.8 to 11.6) and LC-MS/MS of +1.3 % (7.4,-13.3 to 15.8) against their respective method group means. Comparison of measurements in the adult samples showed a strong correlation (r2=0.9331) and concordance (CCC=0.959) between the two methods. LC-MS/MS values were lower than DiaSorin by an overall mean (±SD, 95 %CI) of-1.6 (±14.3,-29.6 to 26.5) pmol/L with an increased negative bias at higher concentrations. In the paediatric samples, weaker correlation (r2=0.6536) and concordance (CCC=0.782) were observed, with greater bias mean (±SD, 95 %CI) of-9.8 (±23.4,-55.7 to 35.9) pmol/L. The variability in the paediatric samples was not associated with concentration or participant age. There was an increase in the correlation and concordance when 1,25(OH)2D2 was included in the analysis. It is likely that the metabolites of vitamin D present in the paediatric population contributed to the measurement of 1,25(OH)2D. The inconsistent agreement highlights the need for better assay harmonisation and paediatric reference intervals using LC-MS/MS method.

Computerized QT and QTc Measurements From Bedside ICU Monitors Are Similar to Those Derived From a Standard 12-Lead ECG.

Evaluate the agreement between computerized QT/QTc measurements from the bedside monitor (four leads) and a time-matched standard 12-lead ECG. Prospective observational study in three adult ICUs. QT/QTc measurements were obtained from a convenience sample, and the two ECG types were ≤ 30 min apart. Agreement was evaluated using Bland-Altman analysis. A total of 120 patients were evaluated for inclusion, and 60 (50%) had a 12-lead ECG for comparison. The mean bias difference for QT measurements was not statistically different (β = -2.47, 95% CI = 5.50 to -11.05; p = 0.44; limits of agreement (LOA) = -64.37 to 59.44). Similar non-statistical differences were observed for QTc (β = -3.20, 95% CI = 5.50 to -11.05; p = 0.44; LOA = -67.43 to 61.03). There was good agreement for both QT and QTc measurements between the two methods. These pilot data are promising and suggest QT/QTc measurements from bedside monitors (four leads) may be an acceptable alternative to obtaining additional standard 12-lead ECGs. Given that half of the ICU patients screened did not have a 12-lead ECG recorded, bedside monitor QT/QTc's could identify at-risk patients. However, an evaluation in a larger sample and non-ICU patients is warranted.

Arctic Sea Ice Surface Temperature Retrieval from FengYun-3A MERSI-I Data

Arctic sea-ice surface temperature (IST) is an important environmental and climatic parameter. Currently, wide-swath sea-ice surface temperature products have a spatial resolution of approximately 1000 m. The Medium Resolution Spectral Imager (MERSI-I) offers a thermal infrared channel with a wide-swath width of 2900 km and a high spatial resolution of 250 m. In this study, we developed an applicable single-channel algorithm to retrieve ISTs from MERSI-I data. The algorithm accounts for the following challenges: (1) the wide range of incidence angle; (2) the unstable snow-covered ice surface; (3) the variation in atmospheric water vapor content; and (4) the unique spectral response function of MERSI-I. We reduced the impact of using a constant emissivity on the IST retrieval accuracy by simulating the directional emissivity. Different ice surface types were used in the simulation, and we recommend the sun crust type as the most suitable for IST retrieval. We estimated the real-time water vapor content using a band ratio method from the MERSI-I near-infrared data. The results show that the retrieved IST was lower than the buoy measurements, with a mean bias and root-mean-square error (RMSE) of −1.928 K and 2.616 K. The retrieved IST is higher than the IceBridge measurements, with a mean bias and RMSE of 1.056 K and 1.760 K. Compared with the original algorithm, the developed algorithm has higher accuracy and reliability. The sensitivity analysis shows that the atmospheric water vapor content with an error of 20% may lead to an IST retrieval error of less than 1.01 K.

Optimal Blood Sampling Time for Area under the Concentration-Time Curve Estimation of Vancomycin by Assessing the Accuracy of Four Bayesian Software.

Effective blood sampling times, beyond trough and peak levels, have not been determined for estimating vancomycin's area under the concentration-time curve (AUC) using the Bayesian software. The aim of this study was to evaluate the accuracy of AUC estimation at different blood sampling times during the same dosing interval at steady state utilizing data from a prior phase I trial of vancomycin. Six healthy adult participants were sampled following intravenous administration of 1 g vancomycin for 1.5 h every 12 h. The AUC was estimated using four software packages and four population pharmacokinetic models. Accuracy was assessed using bias (difference between the estimated and reference AUC) and imprecision (absolute percentage difference between the estimated and reference AUC). The accuracy varied with the sampling time. The optimal two-point sampling times were determined to be 2.5 and 5.5 h post-injection using software packages for EasyTDM, Practical AUC-guided therapeutic drug monitoring (TDM), and Anti-MRSA Agents TDM Analysis Software (incorporating Rodvold, Yamamoto, and Yasuhara models). In these estimations, the mean bias (range, -1.7 to 9.5 µg·h/mL) was unbiased and the mean imprecision (range, -3.0% to 5.0%) was precise. The optimal one-point sampling time was 5.5 h post-injection for Anti-MRSA Agents TDM Analysis Software, which incorporated the Yamamoto and Yasuhara models. In conclusion, optimal blood sampling times may vary depending on the software and model used. Our findings suggest that identifying specific sampling times could improve the efficacy of TDM in clinical practice.

Evaluating the Two-Source Energy Balance Model Using MODIS Data for Estimating Evapotranspiration Time Series on a Regional Scale

Estimating daily continuous evapotranspiration (ET) can significantly enhance the monitoring of crop stress and drought on regional scales, as well as benefit the design of agricultural drought early warning systems. However, there is a need to verify the models’ performance in estimating the spatiotemporal continuity of long-term daily evapotranspiration (ETd) on regional scales due to uncertainties in satellite measurements. In this study, a thermal-based two-surface energy balance (TSEB) model was used concurrently with Terra/Aqua MODIS data and the ERA5 atmospheric reanalysis dataset to calculate the surface energy balance of the soil–canopy–atmosphere continuum and estimate ET at a 1 km spatial resolution from 2000 to 2022. The performance of the model was evaluated using 11 eddy covariance flux towers in various land cover types (i.e., savannas, woody savannas, croplands, evergreen broadleaf forests, and open shrublands), correcting for the energy balance closure (EBC). The Bowen ratio (BR) and residual (RES) methods were used for enforcing the EBC in the EC observations. The modeled ET was evaluated against unclosed ET and closed ET (ETBR and ETRES) under clear-sky and all-sky observations as well as gap-filled data. The results showed that the modeled ET presented a better agreement with closed ET compared to unclosed ET in both Terra and Aqua datasets. Additionally, although the model overestimated ETd across all different land cover types, it successfully captured the spatiotemporal variability in ET. After the gap-filling, the total number of days compared with flux measurements increased substantially, from 13,761 to 19,265 for Terra and from 13,329 to 19,265 for Aqua. The overall mean results including clear-sky and all-sky observations as well as gap-filled data with the Aqua dataset showed the lowest errors with ETRES, by a mean bias error (MBE) of 0.96 mm.day−1, an average mean root square (RMSE) of 1.47 mm.day−1, and a correlation (r) value of 0.51. The equivalent figures for Terra were about 1.06 mm.day−1, 1.60 mm.day−1, and 0.52. Additionally, the result from the gap-filling model indicated small changes compared with the all-sky observations, which demonstrated that the modeling framework remained robust, even with the expanded days. Hence, the presented modeling framework can serve as a pathway for estimating daily remote sensing-based ET on regional scales. Furthermore, in terms of temporal trends, the intra-annual and inter-annual variability in ET can be used as indicators for monitoring crop stress and drought.

Improving Atmospheric Correction Algorithms for Sea Surface Skin Temperature Retrievals from Moderate-Resolution Imaging Spectroradiometer Using Machine Learning Methods

Satellite-retrieved sea-surface skin temperature (SSTskin) is essential for many Near-Real-Time studies. This study aimed to assess the potential to improve the accuracy of satellite-based SSTskin retrieval in the Caribbean region by using atmospheric correction algorithms based on four readily available machine learning (ML) approaches: eXtreme Gradient Boosting (XGBoost), Support Vector Regression (SVR), Random Forest (RF), and the Artificial Neural Network (ANN). The ML models were trained on an extensive dataset comprising in situ SST measurements and atmospheric state parameters obtained from satellite products, reanalyzed datasets, research cruises, surface moorings, and drifting buoys. The benefits and shortcomings of various ML methods were assessed through comparisons with withheld in situ measurements. The results demonstrate that the ML-based algorithms achieve promising accuracy, with mean biases within 0.07 K when compared with the buoy data and ranging from −0.107 K to 0.179 K relative to the ship-derived SSTskin data. Notably, both XGBoost and RF stand out for their superior correlation and efficacy in the statistical results of validation. The improved SSTskin derived using the ML-based algorithms could enhance our understanding of vital oceanic and atmospheric characteristics and have the potential to reduce uncertainty in oceanographic, meteorological, and climate research.

Accuracy, Reproducibility, and Responsiveness to Treatment of Home Spirometry in Cystic Fibrosis: Multicenter, Retrospective, Observational Study.

Portable spirometers are increasingly used to measure lung function at home, but doubts about the accuracy of these devices persist. These doubts stand in the way of the digital transition of chronic respiratory disease care, hence there is a need to address the accuracy of home spirometry in routine care across multiple settings and ages. This study aimed to assess the accuracy, reproducibility, and responsiveness to the treatment of home spirometry in long-term pediatric and adult cystic fibrosis care. This retrospective observational study was carried out in 5 Dutch cystic fibrosis centers. Home spirometry outcomes (forced expiratory volume in one second [FEV1], and forced vital capacity [FVC]) for 601 anonymized users were collected during 3 years. For 81 users, data on clinic spirometry and elexacaftor/tezacaftor/ivacaftor (ETI) use were available. Accuracy was assessed using Bland-Altman plots for paired clinic-home measurements on the same day and within 7 days of each other (nearest neighbor). Intratest reproducibility was assessed using the American Thoracic Society/European Respiratory Society repeatability criteria, the coefficient of variation, and spirometry quality grades. Responsiveness was measured by the percentage change in home spirometry outcomes after the start of ETI. Bland-Altman analysis was performed for 86 same-day clinic-home spirometry pairs and for 263 nearest neighbor clinic-home spirometry pairs (n=81). For both sets and for both FEV1 and FVC, no heteroscedasticity was present and hence the mean bias was expressed as an absolute value. Overall, home spirometry was significantly lower than clinic spirometry (mean ΔFEV1clinic-home 0.13 L, 95% CI 0.10 to 0.19; mean ΔFVCclinic-home 0.20 L, 95% CI 0.14 to 0.25) and remained lower than clinic spirometry independent of age and experience. One-way ANOVA with post hoc comparisons showed significantly lower differences in clinic-home spirometry in adults than in children (Δmean 0.11, 95% CI -0.20 to -0.01) and teenagers (Δmean 0.14, 95% CI -0.25 to -0.02). For reproducibility analyses, 2669 unique measurement days of 311 individuals were included. Overall, 87.3% (2331/2669) of FEV1 measurements and 74.3% (1985/2669) of FVC measurements met reproducibility criteria. Kruskal-Wallis with pairwise comparison demonstrated that for both FVC and FEV1, coefficient of variation was significantly lower in adults than in children and teenagers. A total of 5104 unique home measurements were graded. Grade E was given to 2435 tests as only one home measurement was performed. Of the remaining 2669 tests, 43.8% (1168/2669) and 43.6% (1163/2669) received grade A and B, respectively. The median percentage change in FEV1 from baseline after initiation of ETI was 19.2% after 7-14 days and remained stable thereafter (n=33). Home spirometry is feasible but not equal to clinic spirometry. Home spirometry can confirm whether lung functions remain stable, but the context of measurement and personal trends are more relevant than absolute outcomes.

Modeling of Zenith Tropospheric Delay using ERA5 data over East African region

The Zenith Tropospheric Delay (ZTD) is a crucial parameter in meteorology and climate research, often estimated from surface meteorological parameters and Global Navigation Satellite Systems (GNSS) observations. In East Africa, the lack of reliable surface meteorological data and gaps in GNSS observations compromise the accuracy and reliability of ZTD data. To address this issue, site-specific ZTD models were developed using ERA5 data from 2013 to 2017, employing Empirical Orthogonal Function (EOF) analysis. The accuracy of the proposed EOF models was validated using the tropospheric product from the Nevada Geodetic Laboratory (NGL) as a reference and compared to the Global Pressure and Temperature 3 (GPT3) ZTD model. The results of the study show that the EOF ZTD models significantly outperformed the GPT3 model, reducing Mean bias (MnB) by 72.3% and Root Mean Square Error (RMSE) by 3.0%. EOF models performed particularly well for stations near the equator (latitudes 4°S and 4°N) and between latitudes 12° S and 4° S in terms of MnB and RMSE, respectively. Seasonally, EOF models surpassed the GPT3 model in MnB and RMSE across most seasons near the equator, except during the September-October-November (SON) period, where GPT3 showed an 85.5% better performance in MnB. For stations between latitudes 12° S and 4° S, GPT3 generally performed better in terms of RMSE, except during the March-April-May (MAM) period, where the EOF model excelled. However, the EOF model consistently showed better (reduced) MnB in this region. This study demonstrates that the EOF method is a viable alternative for estimating ZTD in areas with limited surface meteorological data and GNSS observation gaps. The site-specific ZTD models developed using the EOF method can significantly improve the accuracy and reliability of ZTD data, with broad applications in geodesy, atmospheric science, and navigation among others.

Evaluation of a new soluble transferrin receptor assay and comparison to three measurement procedures.

Soluble transferrin receptor (sTfR) is a useful marker in the differentiation of anemia. Clinical utility is limited by lack of standardization between measurement procedures and interpretative recommendations. Our objective was to evaluate the analytical performance of a research sTfR immunoturbidimetric assay (Alinity c, Abbott Diagnostics) and compare it to three established measurement procedures. Assay imprecision was assessed with 7 panels across the analytical measuring interval. 159 patient samples were measured across four instrument systems (Alinity c [Abbott Diagnostics], Tina-quant c502 [Roche Diagnostics], Quantex Biokit [Werfen], and ACCESS [Beckman Coulter]). Ferritin was also measured to calculate an sTfR/Log Ferritin ratio. Sera from 100 reference individuals were assayed for sTfR and ferritin (Alinity) for reference interval (RI) verification (sTfR) or establishment (sTfR index). Assay imprecision met defined goals. Method comparison between Alinity c and ACCESS sTfR assays showed good agreement (slope: 1.06, intercept: -0.12, r: 0.989). Comparisons across other assays demonstrated significant proportional bias with slopes ranging from 0.44 (Tina-quant c502, mean bias: -2.52 mg/L) to 1.24 (Quantex Biokit, mean bias: 0.60 mg/L). A proportional bias was observed between other instruments. While the sTfR RI was verified on the Alinity assay, agreement in interpretation (within vs outside RI) between Alinity and other platforms ranged from 74.2-80.5 %. We report the first characterization of the performance of a research sTfR immunoturbidimetric assay (Alinity c, Abbott Diagnostics). Our findings emphasize the lack of harmonization between measurement procedures and result interpretation for sTfR and sTfR index, necessitating standardization efforts and clinical studies.

Quantitative neuromuscular monitoring with train-of-four ratio using sonomechanomyography (SMMG): Toward more sensitive evaluation of neuromuscular blockade

In the post-anesthesia care unit, there is a high occurrence of residual neuromuscular blockade, which puts patients at risk of negative consequences such as hypoxia. Assessment based on the train-of-four ratio (TOFR) has been used to avoid residual neuromuscular blockade when the TOFR is greater than 0.9, measured at the adductor pollicis muscle (APM). The most commonly used quantitative neuromuscular monitoring (QNM) modalities include acceleromyography (AMG) and electromyography (EMG). However, the poor user-friendliness of current QNM methods hinders their widespread adoption. To overcome this, we developed a new monitoring method using ultra-fast ultrasound imaging to generate a two-dimensional map of muscle transient motion, i.e., sonomechanomyography (SMMG). SMMG of the APM and AMG of the thumb were used to get the TOFR of 20 normal adults. The results showed no significant difference between the left and right hands for both AMG and SMMG TOFR, with p-values larger than 0.05. In addition, the mean accuracy of SMMG TOFR (0.6% relative error) was higher than AMG (1.4% relative error). Moreover, the Bland–Altman plot showed that all the difference values were within the limits of agreement and the mean bias was 0.02, indicating that the two methods had a very good agreement. In particular, using SMMG did not require additional calibration before testing. Overall, the results demonstrated that the method has the potential as a new QNM approach for further clinical studies to benefit patients in need. To demonstrate its clinical potential, further studies are required to evaluate this method in patients during and post-anesthesia.

ICESat-2 data denoising and forest canopy height estimation using Machine Learning

Supervised classification methods can distinguish between noise and signal in ice, cloud, and land elevation satellite-2 (ICESat-2) data across various feature perspectives and autonomously optimize parameters. Nevertheless, model generalization remains a significant limitation for practical applications. This study focuses on developing a universal denoising model for ICESat-2 using machine learning algorithms and analyzing its spatial transferability under various forest and terrain conditions. A photon-denoising feature parameter system is developed based on the analysis of the three-dimensional distribution of photons in forested regions. This system reduces the parameters dependent on absolute physical quantities and increases those that are less influenced by terrain and forest features to enhance the model’s transferability. Subsequently, automated machine learning algorithms (AutoML) are used for model selection and parameter optimization across six non-parametric regression models. We evaluate the accuracies of the local, global, and transfer models in estimating canopy height across four representative forested areas in China. Results show that the algorithm can effectively distinguish between signal and noise photons. The estimated canopy heights from signal photons are highly consistent with heights obtained using airborne laser scanning (ALS), exhibiting a Pearson correlation coefficient (r) of 0.89, root mean square errors (RMSE) of 3.75 m, relative root mean square error (rRMSE) of 0.27, relative bias (rBias) of −0.11 and mean Bias of −1.45 m. Notably, the accuracy of canopy height estimation by the global model has increased by an average of 21 % compared to ICESat-2 land-vegetation along-track products (ATL08). Furthermore, the model exhibits significant spatial transfer capabilities, with the accuracies of the transfer model exceeding those of ATL08 by margins ranging from 4 % to 41 %. This study marks a significant advancement in photon-denoising methodologies, providing a robust and transferable solution for large-scale environmental data analysis.

Assessment of reference evapotranspiration forecasting in the Mediterranean climate of Central Chile using the ASCE standardized Penman-Monteith equation, the Hargreaves-Samani equation, and weather predictions from the Global Forecast System model

Accurate estimation of crop evapotranspiration (ETc) is essential for effective agricultural water resource planning. To determine ETc it is common multiplying the reference evapotranspiration (ETref) by an appropriate crop coefficient (Kc). Forecasting ETref could be particularly beneficial for irrigation scheduling. In this work, daily and cumulative (7 days) ETref predictions were estimated with the Penman-Monteith (PM) and the Hargreaves-Samani (HS) models using environmental variables obtained from the Global Forecast System (GFS). Results were compared with ETref calculated with the ASCE (American Society of Civil Engineers) PM equation from measured daily values of solar radiation, air temperature, relative humidity and wind speed. Four weather stations located in Central Chile during 2020 to 2022 were used. The best results were obtained using the HS equation, for all weather stations, the statistical indicators for daily ETref predictions fall within the following ranges: Accuracy (defined as the percentage of daily ETref predictions with an absolute error within 1mmd−1) varies between 72.8% and 91.8%; the root mean square error (RMSE) ranges from 0.59mmd−1 to 0.98mmd−1; the normalized root mean square error (NRMSE) is between 16.1% and 24.1%; the mean bias error (MBE) is between 0.10mmd−1 and 0.36mmd−1; and the normalized mean bias error (NMBE) varies from 2.5% to 8.3%. On the other hand, the indicators for the 7-day cumulative ETref forecasts were: RMSE from 2.86mm to 5.08mm; NRMSE from 11.0% to 16.3%; MBE from 0.66mm to 2.57mm; and NMBE from 2.4% to 8.3%. These results suggest that HS ETref values predicted from GFS temperature forecasts could be effective for foreseeing near-future water demand and enhancing the accuracy of irrigation scheduling and management in Central Chile.

Agreement between lacrimal fluid and serum for detecting urea nitrogen and creatinine in dogs.

To determine whether urea nitrogen and creatinine levels differ in lacrimal fluid (LF) and serum (SER) in nonazotemic (control) and azotemic dogs and whether there is an agreement between LF and SER. A prospective observational study was performed at the Auburn University Small Animal Teaching Hospital between May 2023 and March 2024. Forty control and 38 azotemic dogs were enrolled. Twenty microliters of LF per eye was collected with microcapillary tubes, and 3 mL of blood was drawn. Bland-Altman plot and intraclass correlation coefficient (ICC) were used to evaluate the agreement between LF and SER. There was good agreement between LF and SER levels of urea nitrogen in the control group (Bland-Altman plot mean bias of -0.8108 ± 2.407 mg/dL; ICC of 0.874 [95% CI, 0.773 to 0.934]) and the azotemic group (Bland-Altman plot mean bias of -9.681 ± 23.89 mg/dL; ICC of 0.82 [95% CI, 0.658 to 0.906]). There was poor agreement between LF and SER concentrations for creatinine in the control and azotemic groups, with only 26 dogs with creatinine detectable in LF. Lacrimal fluid and SER concentrations of urea nitrogen showed good agreement in both the control and azotemic groups, whereas poor agreement was found for creatinine in both groups. Measurement of urea nitrogen in LF may provide an alternative to blood for diagnosing uremia. However, additional research is necessary before substituting LF for SER.

Secondary fastener (tape) placement on a blood pressure cuff causes interindividual Doppler blood pressure measurement variability in healthy large-breed dogs.

The objective of this study was to compare Doppler blood pressure (DBP) measurements between blood pressure cuffs (BPCs) with and without a secondary fastener (tape) in healthy nonanesthetized large-breed dogs. 28 healthy dogs. Superficial palmar venous arch DBP measurements with and without tape on the BPC were performed in lateral recumbency with Doppler ultrasound. Each method was performed 6 consecutive times, with the final 5 values averaged. Bland-Altman plots were made and limits of agreement calculated. The limits of agreement were -40.8 (95% CI, -55.6 to -26.0) to 45.6 (95% CI, 30.8 to 60.4), indicating that the DBP measured with tape would be expected to be between 40.8 mm Hg below and 45.6 mm Hg above that measured without tape 95% of the time. The mean bias estimate was 2.4 mm Hg (SD, 22.0; 95% CI, -6.1 to 11.0; P = .724), indicating that DBP measurements with tape averaged 2.4 mm Hg higher than without tape. Doppler blood pressure measurements obtained with secondary fasteners on the BPCs differed by > 10 mm Hg compared to the BPCs' standard hook-and-loop fasteners 54% (30 of 56) of the time. Blood pressure cuffs with dysfunctional hook-and-loop fasteners should be replaced due to poor clinical reliability of DBP measurements. However, it is unclear whether Doppler sphygmomanometry is more accurate with the use of traditional Velcro BPC fasteners or with tape BPC fasteners.

EVALUATION OF DAILY SATELLITE AND REANALYSIS OF RAINFALL DATA OVER SOUTH SUMATRA REGION

The limitations of good rainfall data due to constraints on direct measurements can be overcome by using satellite data or reanalysis data. The use of this data must, of course, go through a validation process first. This research aims to evaluate daily data from the Tropical Rainfall Measurement Mission (TRMM) version 3B42RT (TRMM_3B42RT) and European Center for Medium-Range Weather Forecasts (ECMWF) Reanalysis 5 (ERA5) data against data collected from 17 rain gauges in the Sumatra region South. Evaluation is carried out based on the Correlation Coefficient (CC), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Bias Error (MBE) values between the data. In addition, the estimation capabilities of TRMM_3B42RT and ERA5 were evaluated based on the Probability of Detection (POD), False Alarm Ratio (FAR), and Critical Success Index (CSI) values. The results show a very high correlation between TRMM_3B42RT and ERA5 with rain gauge data, especially in terms of monthly data. These values (monthly data) for TRMM_3B42RT and ERA5 data are 0.3-0.9 and 0.2-0.9, respectively. The RMSE values of TRMM_3B42RT and ERA5 data in monthly analysis are 75-250mm/month and 100-180mm/month, respectively. The forecasting performance of TRMM_3B42RT and ERA5 shows good results, especially for moderate rainfall in daily data and heavy rainfall in monthly data. The results of this analysis show tha the monthly data TRMM_3B42RT is more in line with the station data and can be used in further research.

Bias evaluation and minimization for estuarine total dissolved solids (TDS) patterns constructed using spatial interpolation techniques

This study evaluated and minimized bias in estuarine total dissolved solids (TDS) patterns predicted using geostatistical approaches. The acquired TDS data was divided into three parts: 29 points (60 %) for predicting TDS patterns with spatial interpolation techniques, 12 points (25 %) for validation and bias correction, and 7 points (37 %) for testing the corrected bias. Inverse Distance Weighting (IDW), Ordinary Kriging (OK), Universal Kriging (UK) and Regression Kriging (RK) were applied to map TDS patterns. R-square and Relative Error of Mean showed significant discrepancies between observed and predicted TDS levels. The Mean field bias (MFB) correction technique was applied to minimize bias in TDS patterns. After correcting bias, the TDS values predicted by IDW, OK, UK and RK at random locations deviated from measured values by −2.85 %, 0.71 %, 4.66 %, and − 6.03 % respectively. At the test locations, these values deviated by −1.45 %, 1.41 %, 2.11 %, and − 2.65 % respectively.

Spectral Estimation of Chlorophyll for Non-Invasive Assessment in Apple Orchards

The main aim of our research was to develop a methodology of chlorophyll content in the leaves of apple trees non-invasive assessment in apple orchards and its adaptation to Early Gold and Golden Reinders based on spectral characteristics of chlorophyll content in the canopy. In each measurement period, 30 samples were collected from each of the two apple cultivars studied. For spectral data collection of leaf samples, an AvaSpec 2048 spectrometer was used in the wavelength range 400–1000 nm in three replicates. Principal component analysis (PCA) with varimax rotation was used to identify the wavelength with the highest factor weight to identify the chlorophyll-sensitive wavelength. The models were calibrated with 2/3 of the values in the database and validated with the remaining 1/3. The simple linear regression method generated the model for estimating chlorophyll. The coefficient of determination (R2) was used to compare the strength of the regression models, and the Root Mean Square Error (RMSE), Normalized Root Mean Square Error (NRMSE), Nash–Sutcliffe efficiency (NSE), Mean Absolute Error (MAE) and Mean Bias Error (MBE) functions were used to measure the accuracy of the estimator models. These metrics help to quickly assess how reliable and accurate a model’s predictions are. Nine indices were obtained based on the precision values, and CHLapple1 performed best (R2 = 0.633, RMSE = 298.28 µg/g, NRMSE = 9.61%, NSE = 0.60, MBE = 84.59, and MAE = 243.39).

Embedded physical constraints in machine learning to enhance vegetation phenology prediction

ABSTRACT Vegetation phenology plays a pivotal role in ecological processes on terrestrial surfaces and the interactions between the biosphere and atmospheric feedback. Current attempts to retrieve vegetation phenology have primarily depended on vegetation indices extracted from satellite remote sensing imagery. These approaches often neglect the underlying physical mechanisms associated with climatic factors, and there is a notable absence of evaluations and comparisons with field-observed phenology inventory data. To address these limitations, this paper proposes an innovative physical constraint neural networks (PCNNs) model that combines machine learning techniques with physical mechanisms to enhance the accuracy of vegetation phenology predictions. By incorporating meteorological variables into a machine learning model and by using the Moderate-Resolution Imaging Spectroradiometer (MODIS) dataset to identify the vegetation phenology of four vegetation types in North America, this study delved into the relationship between vegetation phenology and climate factors as well as its impacts on ecosystems. Our model demonstrated high accuracy compared to machine learning methods without physical mechanisms when validated by field observations from PhenoCam and the USA National Phenology Network (USA-NPN) spanning from 2001 to 2021. The results show that the overall root mean square error (RMSE) with physical constraints is reduced to 12.37 days, higher by 2.6 days than the machine learning method without physical constraints. We compared four vegetation types using different machine learning and traditional rule-based methods, deciduous vegetation (DV) exhibited the most favorable prediction results, with the RMSE and mean bias error (MBE) as low as 5.71 days and 4.06 days in PCNNs model, respectively. This was followed by evergreen needle-leaved forests and mixed forests with RMSE of 12.32 and 13.28 days, respectively. The stressed deciduous vegetation type had the worst prediction result of 19.86 days (RMSE), and the weighted index of agreement (WIA) attained a value of 0.68. These findings suggest that the model embedded with physical constraints significantly boosted prediction accuracy for four common vegetation types, particularly for DV, compared to the unconstrained ML model. It offers valuable insights into the incorporation of physical mechanisms within machine learning models. This research paves the way for substantial advancements in the field of land surface phenology, enabling more accurate and reliable predictions of vegetation phenology in various ecological and climatic contexts.

Variation of Modified Combustion Efficiency and Its Impact on Biomass Burning Emission Estimation in Africa

AbstractBiomass burning (BB) is one of the largest sources of trace gases and primary carbonaceous particles in the global troposphere, posing great impacts on air quality and regional climate. Accurate quantification of BB emissions is vital for assessing its environmental and climate impacts. However, there are still large uncertainties in current BB emission inventories due to poorly characterized emission rates under different combustion states. The fixed emission factor (EF) instead of varying EF associated with different combustion efficiencies may be the reason for the bias. Here, based on satellite‐retrieved carbon dioxide (CO2) and carbon monoxide (CO), the modified combustion efficiency (MCE) is derived for fire‐prone regions in Africa. The monthly and inter‐annual variability of MCE shows a good correlation with meteorological variables such as relative humidity. Therefore, variable EF was established based on its statistical relationship with MCE for different fire‐emitted species. Application of such MCE‐dependent EF in the global climate‐chemistry model can greatly improve the performance of wildfire smoke pollution during the fire season, indicated by an increase of 31% in aerosol optical depth (AOD) and a 50% reduction in normalized mean bias compared with AOD observations. The study elucidates the critical role of meteorology in BB emission estimates and highlights the importance of implementing a dynamic fire emission inventory in response to meteorological conditions.

Atmospheric CO2 column concentration over Iran: Emissions, GOSAT satellite observations, and WRF-GHG model simulations

Regarding global warming and climate change, carbon dioxide (CO2) is one of the most important greenhouse gases. Simulating CO2 gas at hourly/weekly time intervals and desired vertical resolution is challenging due to the coarse horizontal resolution of global models. In this study, both column-averaged CO2 mixing ratio (XCO2) and vertical cross sections of CO2 mixing ratio were simulated by the Weather Research and Forecast Green House gas (WRF-GHG) model at spatial resolutions of 30 and 10 km for the Middle East region as the first domain, and Iran as the second domain. Simulations consider the primary CO2 sources (anthropogenic, biogenic, fire, and oceanic) and the Copernicus Atmosphere Monitoring Service (CAMS) dataset. XCO2 retrieved from GOSAT satellite observations was employed to evaluate the simulation results of the column-averaged CO2 concentrations in February and August 2010. The evaluations showed that the spatiotemporal variability of meteorological variables was well simulated by WRF-GHG with correlation coefficients r of 0.86–0.92, 0.67–0.75, and 0.76–0.82 for temperature, wind, and relative humidity, respectively, during February and August 2010. The evaluations also indicated that the WRF-GHG simulations outperformed the global model TM3, with mean bias error values of − 0.79 and 0.45 PPMV for WRF-GHG in February and August, respectively. The percentage contribution of net CO2 emissions from human activities in Iran was calculated as (38.33 % and 23.70 %) of the total emissions, respectively, with values of 4.4 and 0.85 kg/km2 in each month. The net emissions contributions of biogenic, fire, and oceanic sources were evaluated in February and August, with biogenic emissions contributing (31.901 % and 27.66 %), biogenic absorption contributing (24.07 % and 46.63 %), fire emissions contributing (5.7 % and 2.064 %), and oceanic emissions contributing (3.23 × 10−6 % and 2.23 × 10−6 %). Large-scale circulations and biogenic activity are responsible for the major features of the spatial and seasonal distribution of CO2 in the area. In February, column mixing ratios are higher in more northern latitudes; in August, they are higher to the south. Furthermore, the simulated vertical cross sections show high CO2 mixing ratios in the mid-lower troposphere and northerly/northeasterly advection in February; the vertical profile is inverted in August with high concentrations in the lower stratosphere associated with southwesterly advection. However, the interaction between the synoptic and sub-synoptic features with the topography determines the precise dispersion and distribution of CO2. Despite the negligible emissions in central and eastern Iran, these factors play an important role in the observed concentrations in February and August. In August, the areas between the 120-day low-level monsoon flow of Sistan in eastern Iran, and the westerlies over western/southwestern Iran and the Zagros Mountain range, interact with the heat-low in the Iranian plains and with the Arabian subtropical high. In February, the dominant winds in western Asia were west winds, with the main source of pollution coming from the northeastern regions flowing towards central and eastern Iran. The equatorward displacement of the polar jet stream and the passage of low-pressure systems from the west in winter cause a temporary reduction in CO2 concentrations.