Simultaneous retrieval of sugarcane variables from Sentinel-2 data using Bayesian regularized neural network.

Quantifying biophysical and biochemical vegetation variables is of great importance in precision agriculture. Here, the ability of artificial neural networks (ANNs) to generate multiple outputs is exploited to simultaneously retrieve Leaf area index (LAI), leaf sheath moisture (LSM), leaf chlorophyll content (LCC), and leaf nitrogen concentration (LNC) of sugarcane from Sentinel-2 spectra. We apply a type of ANNs, Bayesian Regularized ANN (BRANN), which incorporates the Bayes' theorem into a regularization scheme to tackle the overfitting problem of ANN and improve its generalizability. Quantitatively assessing the result accuracy indicated RMSE values of 0.48 (m2/m2) for LAI, 2.36 (% wb) for LSM, 5.85 (µg/cm2) for LCC, and 0.23 (%) for LNC, applying simultaneous retrieval. It was demonstrated that simultaneous retrievals of the variables outperformed the individual retrievals. The superiority of the proposed BRANN over a conventional ANN trained with the Levenberg-Marquardt algorithm was confirmed through statistical comparison of their results. The model was applied over the entire Sentinel-2 images to map the considered variables. The maps were probed to qualitatively evaluate the model performance. The results indicated that the retrievals reasonably represent spatial and temporal variations of the variables. Generally, this study demonstrated that the BRANN simultaneous retrieval model can provide faster and more accurate retrievals than those obtained from conventional ANNs and individual retrievals.

Impact assessment of meteorological and environmental parameters on PM2.5 concentrations using remote sensing data and GWR analysis (case study of Tehran).

The PM2.5 as one of the main pollutants in Tehran city has a devastating effect on human health. Knowing the key parameters associated with PM2.5 concentration is essential to take effective actions to reduce the concentration of these particles. This study assesses the relationship between meteorological (humidity, pressure, temperature, precipitation, and wind speed) and environmental parameters (normalize difference vegetation index and land surface temperature of MODIS satellite data) on PM2.5 concentration in Tehran city. The Geographically Weighted Regression (GWR) was employed to assess the impact of key parameters on PM2.5 concentrations in winter and summer. For this purpose, first the seasonal average of meteorological data were extracted and synchronized to satellite data. Then, using the ordinary least square model, the important parameters related to PM2.5 concentration were determined and evaluated. Finally, using the GWR model, the relationships between parameters related to PM2.5 concentration were analyzed. The results of this study indicate that meteorological and environmental parameters in winter season (71%) have a much higher ability to explain PM2.5 concentration than summer season (40%). In winter, PM2.5 concentration has a negative correlation with vegetation at most parts of the study area, a negative correlation with LST in the western and a positive correlation in the eastern part of the study area, a positive correlation with temperature, and a negative correlation with wind speed in the northeastern part of the study area. Precipitation has a positive correlation with PM2.5 concentration in most parts of the study area in both seasons. But, it was investigated in case of higher precipitation (more than 2 mm), PM2.5 concentration decreases. But, there is no negative relationship in any of the dependent parameters with PM2.5 concentration in summer. In this season, the air temperature parameter showed a high correlation with PM2.5 concentration. Also, spatial variations of the local coefficients for all parameters are higher in winter than in summer.

Normalizing land surface temperature for environmental parameters in mountainous and urban areas of a cold semi-arid climate.

Normalization of land surface temperature (LST) relative to environmental factors is of great importance in many scientific studies and applications. The purpose of this study was to develop physical models based on energy balance equations for normalization of satellite derived LST relative to environmental parameters. For this purpose, a set of remote sensing imagery, meteorological and climatic data recorded in synoptic stations, and soil temperatures measured by data loggers were used. For modeling and normalization of LST, a dual-source energy balance model (dual-EB), taking into account two fractions of vegetation and soil, and a triple -source energy balance model (triple-EB), taking into account three fractions of vegetation, soil and built-up land, were proposed with either regional or local optimization strategies. To evaluate and compare the accuracy of different modeling results, correlation coefficients and root mean square difference (RMSE) were computed between modeled LST and LST obtained from satellite imagery, as well as between modeled LST and soil temperature measured by data loggers. Further, the variance of normalized LST values was calculated and analyzed. The results suggested that the use of local optimization strategy increased the accuracy of the normalization of LST, compared to the regional optimization strategy. In addition, no matter the regional or local optimization strategy was employed, the triple-EB model out-performed consistently the dual-EB model for LST normalization. The results show the efficiency of the local triple-EB model to normalize LST relative to environmental parameters. The correlation coefficients were close to zero between all of the environmental parameters and the normalized LST. In other words, normalized LST was completely independent of the environmental parameters considered by this research.

Quantitative assessment of Urmia Lake water using spaceborne multisensor data and 3D modeling.

Preserving aquatic ecosystems and water resources management is crucial in arid and semi-arid regions for anthropogenic reasons and climate change. In recent decades, the water level of the largest lake in Iran, Urmia Lake, has decreased sharply, which has become a major environmental concern in Iran and the region. The efforts to revive the lake concerns the amount of water required for restoration. This study monitored and assessed Urmia Lake status over a period of 30 years (1984 to 2014) using remotely sensed data. A novel method is proposed that generates a lakebed digital elevation model (LBDEM) for Urmia Lake based on time series images from Landsat satellites, water level field measurements, remote sensing techniques, GIS, and 3D modeling. The volume of water required to restore the Lake water level to that of previous years and the ecological water level was calculated based on LBDEM. The results indicate a marked change in the area and volume of the lake from its maximum water level in 1998 to its minimum level in 2014. During this period, 86% of the lake became a salt desert and the volume of the lake water in 2013 was just 0.83% of the 1998 volume. The volume of water required to restore Urmia Lake from benchmark status (in 2014) to ecological water level (1274.10 m) is 12.546 Bm3, excluding evaporation. The results and the proposed method can be used by national and international environmental organizations to monitor and assess the status of Urmia Lake and support them in decision-making.