A Wavelet-based Neural Network Model to Predict Ambient Air Pollutants’ Concentration

The present paper proposes a wavelet based recurrent neural network model to forecast one step ahead hourly, daily mean and daily maximum concentrations of ambient CO, NO₂, NO, O₃, SO₂ and PM_2.5 — the most prevalent air pollutants in urban atmosphere. The time series of each air pollutant has been decomposed into different time-scale components using maximum overlap wavelet transform (MODWT). These time-scale components were made to pass through Elman network. The number of nodes in the network was decided on the basis of the strength (power) of the corresponding input signals. The wavelet network model was then used to obtain one-step ahead forecasts for a period extending from January 2009 to June 2010. The model results for out of sample forecast are reasonably good in terms of model performance parameters such as mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE), normalized mean absolute error (NMSE), index of agreement (IOA) and standard average error (SAE). The MAPE values for daily maximum concentrations of CO, NO₂, NO, O₃, SO₂ and PM2.5 were found to be 9.5%, 17.37%, 21.20%, 13.79%, 17.77% and 11.94%, respectively, at ITO, Delhi, India. Bearing in mind that the forecasts are for daily maximum concentrations tested over a long validation period, the forecast performance of the model may be considered as reasonably good. The model results demonstrate that a judicious selection of wavelet network design may be employed successfully for air quality forecasting.     

Estimation of spatial distrubition of groundwater level and risky areas of seawater intrusion on the coastal region in Çarşamba Plain,Turkey, using different interpolation methods

Groundwater level plays a significant role in coastal plains. Heavy pumping and excessive use of near-coast groundwater can increase the intrusion of seawater into the aquifers. In the present study, groundwater levels were measured at 59 groundwater wells at different times during pre- and post-irrigation seasons (April and September of the year 2012) in Çar?amba Plain, Turkey. To select the best method, two deterministic interpolation methods (inverse distance weighing (IDW) with the weights of 1, 2, and 3 and radial basis function (RBF) with spline with tension (SPT) and completely regularized spline (CRS)) and two stochastic methods (ordinary kriging (OK) with spherical, exponential, and Gaussian variograms) and cokriging (COK)) were compared and then the best interpolation method was used to evaluate the spatial distribution of groundwater levels in different seasons and seasonal changes. A total of nine different techniques were tested. Also, risky areas of seawater intrusion in coastal area were determined using the best methods for two periods. The performance of these interpolation methods is evaluated by using a validation test method. Statistical indices of correlation (R ²), mean absolute error (MAE), and root-mean-square error (RMSE) were used to select and validate the best methods. Comparisons between predicted and observed values indicated RBF as the optimal method for groundwater level estimation in April and September. When the best method RBF and the worst method IDW were compared, significant differences were observed in the spatial distribution of groundwater. Results of the study also revealed that excessive groundwater withdrawals during the post-irrigation season dropped the groundwater levels up to 2.0 m in some sections. With regard to seawater intrusion, 9,103 ha of land area was determined to be highly risky and risky.     

Hybrid Neural Networks and Boosted Regression Tree Models for Predicting Roadside Particulate Matter

This paper examines the application of artificial neural network (ANN) and boosted regression tree (BRT) methods in air quality modelling. The methods were applied to developing air quality models for predicting roadside particle mass concentration (PM₁₀, PM_2.5) and particle number counts (PNC) based on air pollution, traffic and meteorological data from Marylebone Road in London. Elastic net, Lasso and principal components analysis were used as feature selection methods for the ANN models to reduce the number of predictor variables and improve their generalisation. The performance of the ANN with feature selection (ANN hybrid) and the BRT models was evaluated and compared using statistical performance metrics. The performance parameters include root mean square error (RMSE), fraction of prediction within a factor of two of the observation (FAC2), mean bias (MB), mean gross error (MGE), the coefficient of correlation (R) and coefficient of efficiency (CoE) values. The input variables selected by the elastic net produced the best performing ANN models. The ANN hybrid produced models performed only slightly better than the BRT models. The R values of the ANN elastic net and BRT models were 0.96 and 0.95 for PM₁₀, 0.96 and 0.96 for PM_2.5 and 0.89 and 0.87 for PNC, respectively. Their corresponding CoE values were 0.72 and 0.70 for PM₁₀, 0.74 and 0.76 for PM_2.5 and 0.81 and 0.71 for PNC respectively. About 80–99% of all the model predictions are within a factor of two of the observed particle concentrations. The BRT models offer more advantages regarding model interpretation and permit feature selection. Therefore, the study recommends the use of BRT over ANN where the model interpretation is a priority.     

Total suspended matter observation in the Pearl River estuary from in situ and MERIS data

Based on the cruise data collected in the Pearl River estuary (PRE) in May 2008, an empirical two-band model by using the ratio of R _rs at 629 and 671 nm was established to retrieve total suspended matter (TSM) concentration with the determination coefficient (R²) of 0.854, mean relative error (MRE) of 7.483%, and root-mean-square error (RMSE) of 1.295 mg L^???1. To match with medium resolution imaging spectrometer (MERIS) bands, in situ remote sensing reflectance was re-sampled to the bandwidth of 10 nm. The relationship between TSM and re-sampled R _rs at 620 nm (MERIS band 6) and 665 nm (MERIS band 7) are obtained (R² = 0.748, RMSE = 1.697 mg L^???1, MRE = 8.785%, n = 13). Additionally, to map the spatial distribution of TSM in the PRE, MERIS level_1B data were calibrated using a multiple linear regression model based on in situ R _rs. Another dataset collected in the PRE in January 2004 was used to validate the two-band model and also applied to map TSM distribution from MERIS image. The comparison between measured TSM values and modeled ones showed satisfactory results (R² = 0.753, MRE = 22.199%, and RMSE = 2.603 mg L^???1).     

Objective approach for analysis of noise source characteristics and acoustic conditions in noisy computerized embroidery workrooms

It is highly important to analyze the acoustic properties of workrooms in order to identify best noise control measures from the standpoint of noise exposure limits. Due to the fact that sound pressure is dependent upon environments, it cannot be a suitable parameter for determining the share of workroom acoustic characteristics in producing noise pollution. This paper aims to empirically analyze noise source characteristics and acoustic properties of noisy embroidery workrooms based on special parameters. In this regard, reverberation time as the special room acoustic parameter in 30 workrooms was measured based on ISO 3382-2. Sound power quantity of embroidery machines was also determined based on ISO 9614-3. Multiple linear regression was employed for predicting reverberation time based on acoustic features of the workrooms using MATLAB software. The results showed that the measured reverberation times in most of the workrooms were approximately within the ranges recommended by ISO 11690-1. Similarity between reverberation time values calculated by the Sabine formula and measured values was relatively poor (R ²?=?0.39). This can be due to the inaccurate estimation of the acoustic influence of furniture and formula preconditions. Therefore, this value cannot be considered representative of an actual acoustic room. However, the prediction performance of the regression method with root mean square error (RMSE)?=?0.23 s and R ²?=?0.69 is relatively acceptable. Because the sound power of the embroidery machines was relatively high, these sources get the highest priority when it comes to applying noise controls. Finally, an objective approach for the determination of the share of workroom acoustic characteristics in producing noise could facilitate the identification of cost-effective noise controls.     

Temporal changes of soil respiration under different tree species

Soil respiration rates were measured monthly (from April 2007 to March 2008) under four adjacent coniferous plantation sites [Oriental spruce (Picea orientalis L.), Austrian pine (Pinus nigra Arnold), Turkish fir (Abies bornmulleriana L.), and Scots pine (Pinus sylvestris L.)] and adjacent natural Sessile oak forest (Quercus petraea L.) in Belgrad Forest—Istanbul/Turkey. Also, soil moisture, soil temperature, and fine root biomass were determined to identify the underlying environmental variables among sites which are most likely causing differences in soil respiration. Mean annual soil moisture was determined to be between 6.3 % and 8.1 %, and mean annual temperature ranged from 13.0°C to 14.2°C under all species. Mean annual fine root biomass changed between 368.09 g/m² and 883.71 g/m² indicating significant differences among species. Except May 2007, monthly soil respiration rates show significantly difference among species. However, focusing on tree species, differences of mean annual respiration rates did not differ significantly. Mean annual soil respiration ranged from 0.56 to 1.09 g?C/m²/day. The highest rates of soil respiration reached on autumn months and the lowest rates were determined on summer season. Soil temperature, soil moisture, and fine root biomass explain mean annual soil respiration rates at the highest under Austrian pine (R ²?=?0.562) and the lowest (R ²?=?0.223) under Turkish fir.     

Modeling seasonal variations of long-term soil CO<Subscript>2</Subscript> emissions in an orchard plantation in a semiarid area,SE Turkey

Emissions of soil CO₂ under different management systems have a significant effect on the carbon balance in the atmosphere. Soil CO₂ emissions were measured from an apricot orchard at two different locations: under the crown of trees (CO₂-UC) and between tree rows (CO₂-BR). For comparison, one other measurement was performed on bare soil (CO₂-BS) located next to the orchard field. Analytical data were obtained weekly during 8 years from April 2008 to December 2016. Various environmental parameters such as air temperature, soil temperature at different depths, soil moisture, rainfall, and relative humidity were used for modeling and estimating the long-term seasonal variations in soil CO₂ emissions using two different methods: generalized linear model (GLM) and artificial neural network (ANN). Before modeling, data were randomly split into two parts, one for calibration and the second for validation, with a varying number of samples in each part. Performances of the models were compared and evaluated using means absolute of estimations (MAE), square root of mean of prediction (RMSEP), and coefficient of determination (R²) values. CO₂-UC, CO₂-BR, and CO₂-BS values ranged from 11 to 3985, from 9 to 2365, and from 8 to 1722 kg ha^?1 week^?1, respectively. Soil CO₂ emissions were significantly correlated (p?<?0.05) with some environmental variables. The results showed that GLM and ANN models provided similar accuracies in modeling and estimating soil CO₂ emissions, as the number of samples in the validation data set increased. The ANN was more advantageous than GLM models by providing a better fit between actual observations and predictions and lower RMSEP and MAE values. The results suggested that the success of environmental variables for estimations of CO₂ emissions using the two methods was moderate.     

Mapping aboveground woody biomass using forest inventory,remote sensing and geostatistical techniques

Mapping forest biomass is fundamental for estimating CO₂ emissions, and planning and monitoring of forests and ecosystem productivity. The present study attempted to map aboveground woody biomass (AGWB) integrating forest inventory, remote sensing and geostatistical techniques, viz., direct radiometric relationships (DRR), k-nearest neighbours (k-NN) and cokriging (CoK) and to evaluate their accuracy. A part of the Timli Forest Range of Kalsi Soil and Water Conservation Division, Uttarakhand, India was selected for the present study. Stratified random sampling was used to collect biophysical data from 36 sample plots of 0.1 ha (31.62 m?×?31.62 m) size. Species-specific volumetric equations were used for calculating volume and multiplied by specific gravity to get biomass. Three forest-type density classes, viz. 10–40, 40–70 and >70 % of Shorea robusta forest and four non-forest classes were delineated using on-screen visual interpretation of IRS P6 LISS-III data of December 2012. The volume in different strata of forest-type density ranged from 189.84 to 484.36 m³ ha^?1. The total growing stock of the forest was found to be 2,024,652.88 m³. The AGWB ranged from 143 to 421 Mgha^?1. Spectral bands and vegetation indices were used as independent variables and biomass as dependent variable for DRR, k-NN and CoK. After validation and comparison, k-NN method of Mahalanobis distance (root mean square error (RMSE)?=?42.25 Mgha^?1) was found to be the best method followed by fuzzy distance and Euclidean distance with RMSE of 44.23 and 45.13 Mgha^?1 respectively. DRR was found to be the least accurate method with RMSE of 67.17 Mgha^?1. The study highlighted the potential of integrating of forest inventory, remote sensing and geostatistical techniques for forest biomass mapping.     

A glimpse at short-term controls of evapotranspiration along the southern slopes of Kilimanjaro

Future climate characteristics of the southern Kilimanjaro region, Tanzania, are mainly determined by local land-use and global climate change. Reinforcing increasing dryness throughout the twentieth century, ongoing land transformation processes emphasize the need for a proper understanding of the regional-scale water budget and possible implications on related ecosystem functioning and services. Here, we present an analysis of scintillometer-based evapotranspiration (ET) covering seven distinct habitat types across a massive climate gradient from the colline savanna woodlands to the upper-mountain Helichrysum zone (940 to 3960 m.a.s.l.). Random forest-based mean variable importance indicates an outstanding significance of net radiation (R _net) on the observed ET across all elevation levels. Accordingly, topography and frequent cloud/fog events have a dampening effect at high elevations, whereas no such constraints affect the energy and moisture-rich submontane coffee/grassland level. By contrast, long-term moisture availability is likely to impose restrictions upon evapotranspirative net water loss in savanna, which particularly applies to the pronounced dry season. At plot scale, ET can thereby be approximated reasonably using R _net, soil heat flux, and to a lesser degree, vapor pressure deficit and rainfall as predictor variables (R ² 0.59 to 1.00). While multivariate regression based on pooled meteorological data from all plots proves itself useful for predicting hourly ET rates across a broader range of ecosystems (R ² = 0.71), additional gains in explained variance can be achieved when vegetation characteristics as seen from the NDVI are considered (R ² = 0.87). To sum up, our results indicate that valuable insights into land cover-specific ET dynamics, including underlying drivers, may be derived even from explicitly short-term measurements in an ecologically highly diverse landscape.     

Analysis of agricultural drought using vegetation temperature condition index (VTCI) from Terra/MODIS satellite data

The most commonly used normalized difference vegetation index (NDVI) from remote sensing often fall short in real-time drought monitoring due to a lagged vegetation response to drought. Therefore, research recently emphasized on the use of combination of surface temperature and NDVI which provides vegetation and moisture conditions simultaneously. Since drought stress effects on agriculture are closely linked to actual evapotranspiration, we used a vegetation temperature condition index (VTCI) which is more closely related to crop water status and holds a key place in real-time drought monitoring and assessment. In this study, NDVI and land surface temperature (T _s) from MODIS 8-day composite data during cloud-free period (September–October) were adopted to construct an NDVI–T _s space, from which the VTCI was computed. The crop moisture index (based on estimates of potential evapotranspiration and soil moisture depletion) was calculated to represent soil moisture stress on weekly basis for 20 weather monitoring stations. Correlation and regression analysis were attempted to relate VTCI with crop moisture status and crop performance. VTCI was found to accurately access the degree and spatial extent of drought stress in all years (2000, 2002, and 2004). The temporal variation of VTCI also provides drought pattern changes over space and time. Results showed significant and positive relations between CMI (crop moisture index) and VTCI observed particularly during prominent drought periods which proved VTCI as an ideal index to monitor terminal drought at regional scale. VTCI had significant positive relationship with yield but weakly related to crop anomalies. Duration of terminal drought stress derived from VTCI has a significant negative relationship with yields of major grain and oilseeds crops, particularly, groundnut.     

The effectiveness of digital soil mapping to predict soil properties over low-relief areas

This study investigates the ability of different digital soil mapping (DSM) approaches to predict some of physical and chemical topsoil properties in the Shahrekord plain of Chaharmahal-Va-Bakhtiari province, Iran. According to a semi-detailed soil survey, 120 soil samples were collected from 0 to 30 cm depth with approximate distance of 750 m. Particle size distribution, coarse fragments (CFs), electrical conductivity (EC), pH, organic carbon (OC), and calcium carbonate equivalent (CCE) were determined. Four machine learning techniques, namely, artificial neural networks (ANNs), boosted regression tree (BRT), generalized linear model (GLM), and multiple linear regression (MLR), were used to identify the relationship between soil properties and auxiliary information (terrain attributes, remote sensing indices, geology map, existing soil map, and geomorphology map). Root-mean-square error (RMSE) and mean error (ME) were considered to determine the performance of the models. Among the studied models, GLM showed the highest performance to predict pH, EC, clay, silt, sand, and CCE, whereas the best model is not necessarily able to make accurate estimation. According to RMSE%, DSM has a good efficiency to predict soil properties with low and moderate variabilities. Terrain attributes were the main predictors among different studied auxiliary information. The accuracy of the estimations with more observations is recommended to give a better understanding about the performance of DSM approach over low-relief areas.     

Monitoring and assessment of seasonal land cover changes using remote sensing: a 30-year (1987–2016) case study of Hamoun Wetland,Iran

The availability of Landsat data allows improving the monitoring and assessment of large-scale areas with land cover changes in rapid developing regions. Thus, we pretend to show a combined methodology to assess land cover changes (LCCs) in the Hamoun Wetland region (Iran) over a period of 30-year (1987–2016) and to quantify seasonal and decadal landscape and land use variabilities. Using the pixel-based change detection (PBCD) and the post-classification comparison (PCC), four land cover classes were compared among spring, summer, and fall seasons. Our findings showed for the water class a higher correlation between spring and summer (R²?=?0.94) than fall and spring (R²?=?0.58) seasons. Before 2000, ~?50% of the total area was covered by bare soil and 40% by water. However, after 2000, more than 70% of wetland was transformed into bare soils. The results of the long-term monitoring period showed that fall season was the most representative time to show the inter-annual variability of LCCs monitoring and the least affected by seasonal-scale climatic variations. In the Hamoun Wetland region, land cover was highly controlled by changes in surface water, which in turn responded to both climatic and anthropogenic impacts. We were able to divide the water budget monitoring into three different ecological regimes: (1) a period of high water level, which sustained healthy extensive plant life, and approximately 40% of the total surface water was retained until the end of the hydrological year; (2) a period of drought during high evaporation rates was observed, and a mean wetland surface of about 85% was characterized by bare land; and (3) a recovery period in which water levels were overall rising, but they are not maintained from year to year. After a spring flood, in 2006 and 2013, grassland reached the highest extensions, covering till more than 20% of the region, and the dynamics of the ecosystem were affected by the differences in moisture. The Hamoun wetland region served as an important example and demonstration of the feedbacks between land cover and land uses, particularly as pertaining to water resources available to a rapidly expanding population.     

Spatial and temporal estimation of soil loss for the sustainable management of a wet semi-arid watershed cluster

The ungauged wet semi-arid watershed cluster, Seethagondi, lies in the Adilabad district of Telangana in India and is prone to severe erosion and water scarcity. The runoff and soil loss data at watershed, catchment, and field level are necessary for planning soil and water conservation interventions. In this study, an attempt was made to develop a spatial soil loss estimation model for Seethagondi cluster using RUSLE coupled with ARCGIS and was used to estimate the soil loss spatially and temporally. The daily rainfall data of Aphrodite for the period from 1951 to 2007 was used, and the annual rainfall varied from 508 to 1351 mm with a mean annual rainfall of 950 mm and a mean erosivity of 6789 MJ mm ha^?1 h^?1 year^?1. Considerable variation in land use land cover especially in crop land and fallow land was observed during normal and drought years, and corresponding variation in the erosivity, C factor, and soil loss was also noted. The mean value of C factor derived from NDVI for crop land was 0.42 and 0.22 in normal year and drought years, respectively. The topography is undulating and major portion of the cluster has slope less than 10°, and 85.3 % of the cluster has soil loss below 20 t ha^?1 year^?1. The soil loss from crop land varied from 2.9 to 3.6 t ha^?1 year^?1 in low rainfall years to 31.8 to 34.7 t ha^?1 year^?1 in high rainfall years with a mean annual soil loss of 12.2 t ha^?1 year^?1. The soil loss from crop land was higher in the month of August with an annual soil loss of 13.1 and 2.9 t ha^?1 year^?1 in normal and drought year, respectively. Based on the soil loss in a normal year, the interventions recommended for 85.3 % of area of the watershed includes agronomic measures such as contour cultivation, graded bunds, strip cropping, mixed cropping, crop rotations, mulching, summer plowing, vegetative bunds, agri-horticultural system, and management practices such as broad bed furrow, raised sunken beds, and harvesting available water using farm ponds and percolation tanks. This methodology can be adopted for estimating the soil loss from similar ungauged watersheds with deficient data and for planning suitable soil and water conservation interventions for the sustainable management of the watersheds.     

Estimation of underwater visibility in coastal and inland waters using remote sensing data

An optical method is developed to estimate water transparency (or underwater visibility) in terms of Secchi depth (Z _sd ), which follows the remote sensing and contrast transmittance theory. The major factors governing the variation in Z _sd , namely, turbidity and length attenuation coefficient (1/(c + K _d ), c = beam attenuation coefficient; K _d  = diffuse attenuation coefficient at 531 nm), are obtained based on band rationing techniques. It was found that the band ratio of remote sensing reflectance (expressed as (R _rs (443) + R _rs (490))/(R _rs (555) + R _rs (670)) contains essential information about the water column optical properties and thereby positively correlates to turbidity. The beam attenuation coefficient (c) at 531 nm is obtained by a linear relationship with turbidity. To derive the vertical diffuse attenuation coefficient (K _d ) at 531 nm, K _d (490) is estimated as a function of reflectance ratio (R _rs (670)/R _rs (490)), which provides the bio-optical link between chlorophyll concentration and K _d (531). The present algorithm was applied to MODIS-Aqua images, and the results were evaluated by matchup comparisons between the remotely estimated Z _sd and in situ Z _sd in coastal waters off Point Calimere and its adjoining regions on the southeast coast of India. The results showed the pattern of increasing Z _sd from shallow turbid waters to deep clear waters. The statistical evaluation of the results showed that the percent mean relative error between the MODIS-Aqua-derived Z _sd and in situ Z _sd values was within ±25%. A close agreement achieved in spatial contours of MODIS-Aqua-derived Z _sd and in situ Z _sd for the month of January 2014 and August 2013 promises the model capability to yield accurate estimates of Z _sd in coastal, estuarine, and inland waters. The spatial contours have been included to provide the best data visualization of the measured, modeled (in situ), and satellite-derived Z _sd products. The modeled and satellite-derived Z _sd values were compared with measurement data which yielded RMSE = 0.079, MRE = ?0.016, and R ²  = 0.95 for the modeled Z _sd and RMSE = 0.075, MRE = 0.020, and R ²  = 0.95 for the satellite-derived Z _sd products.     

Predicting field capacity,wilting point,and the other physical properties of soils using hyperspectral reflectance spectroscopy: two different statistical approaches

In this study, we examined the ability of reflectance spectroscopy to predict some of the most important soil parameters for irrigation such as field capacity (FC), wilting point (WP), clay, sand, and silt content. FC and WP were determined for 305 soil samples. In addition to these soil analyses, clay, silt, and sand contents of 145 soil samples were detected. Raw spectral reflectance (raw) of these soil samples, between 350 and 2,500-nm wavelengths, was measured. In addition, first order derivatives of the reflectance (first) were calculated. Two different statistical approaches were used in detecting soil properties from hyperspectral data. Models were evaluated using the correlation of coefficient (r), coefficient of determination (R ²), root mean square error (RMSE), and residual prediction deviation (RPD). In the first method, two appropriate wavelengths were selected for raw reflectance and first derivative separately for each soil property. Selection of wavelengths was carried out based on the highest positive and negative correlations between soil property and raw reflectance or first order derivatives. By means of detected wavelengths, new combinations for each soil property were calculated using rationing, differencing, normalized differencing, and multiple regression techniques. Of these techniques, multiple regression provided the best correlation (P?<?0.01) for selected wavelengths and all soil properties. To estimate FC, WP, clay, sand, and silt, multiple regression equations based on first(2,310)-first(2,360), first(2,310)-first(2,360), first(2,240)-first(1,320), first(2,240)-first(1,330), and raw(2,260)-raw(360) were used. Partial least square regression (PLSR) was performed as the second method. Raw reflectance was a better predictor of WP and FC, whereas first order derivative was a better predictor of clay, sand, and silt content. According to RPD values, statistically excellent predictions were obtained for FC (2.18), and estimations for WP (2.0), clay (1.8), and silt (1.63) were acceptable. However, sand values were poorly predicted (RDP?=?0.63). In conclusion, both of the methods examined here offer quick and inexpensive means of predicting soil properties using spectral reflectance data.     

Monitoring diel dissolved oxygen dynamics through integrating wavelet denoising and temporal neural networks

Diel dissolved oxygen (DO) time series measured continuously using proximal sensors in situ for a temperate lake were denoised using discrete wavelet transform (DWT) with the orthogonal wavelet families of coiflet, daubechies, and symmlet with order of 10. Diel DO time series denoised were modeled using nine temporal artificial neural networks (ANNs) as a function of water level, water temperature, electrical conductivity, pH, day of year, and hour. Our results showed that time-lag recurrent network (TLRN) using denoised data emulated diel DO dynamics better than the best-performing TLRN using the original data, time-delay neural network (TDNN), and recurrent network (RNN). Daubechies basis dealt with diel DO data slightly better than the other bases given its coefficient of determination (r ²?=?87.1 %), while symmlet performed slightly better than the other bases in terms of root mean square error (RMSE?=?1.2 ppm) and mean absolute error (MAE?=?0.9 ppm).     

Estimation of soil specific surface area using some mechanical properties of soil by artificial neural networks

Soil specific surface area (SSA) is an important property of soil. Depending on the measurement techniques, determination of the SSA is costly and time consuming. Hence, a limited number of studies have been conducted to predict the SSA from the soil variables. In this study, the soil samples were taken from the literature. Fractal parameters (FP) were calculated by the model of Bird et al. (European Journal of Soil Science 51, 55–63, 2000) used as the input variables to predict the SSA. Some studies have been carried out on the prediction capability of the different parameters using the artificial neural networks (ANNs). The ANNs were further used and 20 models were developed to investigate the value of input variables to predict the SSA. The results showed that the PTF13 (RMSE?=?0.13) and PTF18 (RMSE?=?0.13) with the input variables of particle-size distribution and Atterberg limits revealed better performance than the other PTFs (in the training step). It is because of the fact that free swelling index (FSI) and Atterberg limits were closely correlated to the soil clay mineralogy as one of the important factors controlling the SSA. In general, this results demonstrated that the PTF9 with the variables of sand, clay, plastic limit (PL), liquid limit (LL), and FSI showed the best (RMSE?=?0.37) results in the estimation of the SSA. In conclusion, there was not a strong correlation between the soil mechanical properties and SSA but also ANNs were a suitable method to predict the SSA from the soil variables.     

Modified grey model for estimating traffic tunnel air quality

This study compared three forecasting models based on the mean absolute percentage errors (MAPE) of their accuracy in forecasting air pollution in a traffic tunnel: the Grey model (GM), the combination model used four sample point and five sample point prediction with GM (1,1)(GM(1,1)^4 + 5), and the modified grey model (MGM). An MGM was combined using the four points of the original sequence using the original grey prediction GM (1,1) for short-term forecasting. The proposed method cannot only enhance the prediction accuracy of the original grey model, but can also solve the jump data forecasting problem something for which the original grey model is inappropriate. The MAPE was applied to the models, and the MGM found the proposed method to be simple and efficient. The MAPE of MGM, calculated over 3 h of forecasts, were as follows: 10.12 (Upwind), 10.07 (Middle) and 7.68 (Downwind) for CO; 10.79 (Upwind), 6.05 (Middle) and 5.98 (Downwind) for NO _x , and 11.67 (Upwind), 7.32 (Middle) and 4.56 (Downwind) for NMHC. The MGM model results reveal that the combined forecasts can significantly decrease the overall forecasting error. Results of this demonstrate that MGM can accurately forecast air pollution in the Kaohsiung Chung–Cheng Tunnel.     

Investigation of polar organic chemical integrative sampler (POCIS) flow rate dependence for munition constituents in underwater environments

Munition constituents (MC) are present in aquatic environments throughout the world. Potential for fluctuating release with low residence times may cause concentrations of MC to vary widely over time at contaminated sites. Recently, polar organic chemical integrative samplers (POCIS) have been demonstrated to be valuable tools for the environmental exposure assessment of MC in water. Flow rate is known to influence sampling by POCIS. Because POCIS sampling rates (R_s) for MC have only been determined under quasi-static conditions, the present study evaluated the uptake of 2,4,6-trinitrotoluene (TNT), RDX (hexahydro-1,3,5-trinitro-1,3,5-triazine), and 2,4- and 2,6-dinitrotoluenes (DNT), by POCIS in a controlled water flume at 7, 15, and 30 cm/s in 10-day experiments using samplers both within and without a protective cage. Sampling rate increased with flow rate for all MC investigated, but flow rate had the strongest impact on TNT and the weakest impact on RDX. For uncaged POCIS, mean R_s for 30 cm/s was significantly higher than that for 7 cm by 2.7, 1.9, 1.9, and 1.3 folds for TNT, 2,4-DNT, 2,6-DNT, and RDX, respectively. For all MC except RDX, mean R_s for caged POCIS at 7 cm/s were significantly lower than for uncaged samplers and similar to those measured at quasi-static condition, but except for 2,6-DNT, no caging effect was measured at the highest flow rate, indicating that the impact of caging on R_s is flow rate-dependent. When flow rates are known, flow rate-specific R_s should be used for generating POCIS-derived time-averaged concentrations of MC at contaminated sites.     

Large-scale carbon stock assessment of woody vegetation in tropical dry deciduous forest of Sathanur reserve forest,Eastern Ghats,India

Tropical dry forests are one of the most widely distributed ecosystems in tropics, which remain neglected in research, especially in the Eastern Ghats. Therefore, the present study was aimed to quantify the carbon storage in woody vegetation (trees and lianas) on large scale (30, 1 ha plots) in the dry deciduous forest of Sathanur reserve forest of Eastern Ghats. Biomass of adult (≥10 cm DBH) trees was estimated by species-specific allometric equations using diameter and wood density of species whereas in juvenile tree population and lianas, their respective general allometric equations were used to estimate the biomass. The fractional value 0.4453 was used to convert dry biomass into carbon in woody vegetation of tropical dry forest. The mean aboveground biomass value of juvenile tree population was 1.86 Mg/ha. The aboveground biomass of adult trees ranged from 64.81 to 624.96 Mg/ha with a mean of 245.90 Mg/ha. The mean aboveground biomass value of lianas was 7.98 Mg/ha. The total biomass of woody vegetation (adult trees + juvenile population of trees + lianas) ranged from 85.02 to 723.46 Mg/ha, with a mean value of 295.04 Mg/ha. Total carbon accumulated in woody vegetation in tropical dry deciduous forest ranged from 37.86 to 322.16 Mg/ha with a mean value of 131.38 Mg/ha. Adult trees accumulated 94.81% of woody biomass carbon followed by lianas (3.99%) and juvenile population of trees (1.20%). Albizia amara has the greatest biomass and carbon stock (58.31%) among trees except for two plots (24 and 25) where Chloroxylon swietenia contributed more to biomass and carbon stock. Similarly, Albizia amara (52.4%) showed greater carbon storage in juvenile population of trees followed by Chloroxylon swietenia (21.9%). Pterolobium hexapetalum (38.86%) showed a greater accumulation of carbon in liana species followed by Combretum albidum (33.04%). Even though, all the study plots are located within 10 km radius, they show a significant spatial variation among them in terms of biomass and carbon stocks which could be attributed to variation in anthropogenic pressures among the plots as well as to changes in tree density across landscapes. Total basal area of woody vegetation showed a significant positive (R ² = 0.978; P = 0.000) relationship with carbon storage while juvenile tree basal area showed the negative relationship (R ² = 0.4804; P = 0.000) with woody carbon storage. The present study generates a large-scale baseline data of dry deciduous forest carbon stock, which would facilitate carbon stock assessment at a national level as well as to understand its contribution on a global scale.