This research proposes an ensemble method for synergistically combining multiple empirical algorithms to better estimate chlorophyll-a (Chl-a) concentration. In previous studies, different empirical algorithms have been employed separately and a single algorithm was often identified as the most suitable predictor for Chl-a retrieval. Our ensemble method combines different individual algorithms to form an ensemble predictor that exploits advantages of each individual algorithm to maximize the overall estimation accuracy. We evaluated two ensemble predictors: the optimally weighted ensemble predictor and the spectral space partition guided ensemble predictor. The ensemble method has been successfully applied to a Sentinel-2A multispectral image acquired over Harsha Lake, Ohio in 2016. Based on in situ water reference data and satellite imagery, we constructed two ensemble predictors that consist of three individual empirical algorithms/estimators, including 2BDA (two-band algorithm), 3BDA (three-band algorithm), and NDCI (Normalized Difference Chlorophyll Index). For the optimally weighted ensemble predictor, the weights for individual algorithms are computed by solving an overdetermined linear system with the pseudoinverse technique. For the spectral space partition guided ensemble predictor, the rules for partitioning spectral space into spectral regions were established as a decision-tree using the CART method. The optimal Chl-a estimate for a pixel is obtained by selectively using the empirical algorithm in the ensemble that has the highest expected accuracy in the spectral region where the pixel is located. Our assessments suggest that the spectral space partition guided ensemble method performs significantly better than three individual empirical algorithms and also better than the optimally weighted ensemble method. 相似文献
Dam cascades have been a major threat to fishes of the upper Yangtze River. The remaining lotic river segment of the upper mainstem between the Xiangjiaba Dam (XJD) and the Three Gorges Reservoir (TGR) may serve as a critical refuge for endemic fishes. We investigated distribution of ichthyoplankton Botia superciliaris, an endemic species, at three sections preimpoundment and postimpoundment of the XJD, that is, Yibin (close to the XJD), Zhuyang (in the middle of the segment), and Mudong (in the tail of the TGR). Preimpoundment of the XJD, larvae occurred in all three sections, and their abundance tended to be highest at Yibin and lowest at Mudong. Postimpoundment, larval abundance dramatically decreased in all three sections. In particular, no larvae were found at Yibin, and larval abundance tended to be higher at Zhuyang than at Mudong. Eggs were collected from all three sections, and those from Zhuyang accounted for over 90% of the total. Initiation of egg occurrence at Yibin was approximately 1 month later than at Zhuyang and Mudong. Water temperature was lower, and water transparency was higher at Yibin than at Zhuyang and Mudong postimpoundment. We suggest that the decrease in abundance and delayed presence of ichthyoplankton at Yibin postimpoundment of the XJD reflected the impact of upstream dam discharge and that the lower abundance of ichthyoplankton at Mudong compared with Zhuyang both preimpoundment and postimpoundment of the XJD reflected the influence of inundation by the TGR. We recommend that the river sections around Zhuyang become high‐priority conservation areas 相似文献
Quantification of the uncertainty associated with stormwater models should be analyzed before using modelling results to make decisions on urban stormwater control and management programs. In this study, the InfoWorks Integrated Catchment Modelling (ICM) rainfall-runoff model was used to simulate hydrographs at the outfall of a catchment (drainage area 8.3 ha, with 95% pervious areas) in Shenzhen, China. The model was calibrated and validated for two rainfall events with Nash-Sutcliffe efficiency >0.81. The influence of rainfall, model parameters and routing methods on outflow hydrograph of the catchment was systematically studied. The influence of rainfall was analyzed using generated rainfall distributions with random errors and systematic errors (± 30% offsets). Random errors had less influence than systematic errors on peak flow and runoff volume, especially for two rainfall events with larger depths and longer durations. The Monte Carlo simulations using 500 parameter sets were used to verify the equifinality of the nine model parameters and determine the prediction uncertainty. Most of the monitored flows were within the uncertainty range. The influence of two routing methods from rainfall excess to hydrograph was studied. The InfoWorks ICM model incorporating double quasilinear reservoir routing was found to have a larger effect on the simulated hydrographs for rainfall events having larger depths and longer durations than using the U.S. EPA’s Storm Water Management Model nonlinear reservoir routing method did.
At the coastline of the Carey Island, mangroves provide natural protection against the wind-driven coastal waves. The area is located at the west Malaysia within the waters of the Straits of Malacca. Recently, its coastline has been exposed to increasing rates of coastal erosion due to mangrove deforestation. In order to provide mitigating measures, it is necessary to study wave characteristics in this region. For this purpose, we collected 5 years (2009 to 2013) of hourly measurements for wind direction, wave height, wind speed and wave period. Moreover, we used the adaptive neuro-fuzzy inference system (ANFIS) to estimate the wave period and height. The model was trained using the measured data. The validation of the model gave satisfactory R2 values of 0.8484 and 0.9496 for wave height and wave period, respectively. The findings from this study suggest that fuzzy logic based technique satisfactorily predicts the differences between multiple inputs and single output in terms of non-linear relationship. The developed model can be used to further study the effect of non-linear wind-driven waves on the depleting coastal mangrove forests in similar tropical and sub-tropical areas. We suggest further research to test the model in different geographical locations, such as in deep-ocean, narrow straits and other coastal sites, which were not covered in this study. 相似文献