基于无人机高光谱遥感的水体CHL-a反演

无人机平台搭载高光谱传感器获取的高光谱数据具有高分辨率、波谱丰富的特点,本研究基于实测采样点光谱特征、采样点CHL-a浓度,采用相关性分析的方法选取最优波段,构建水体CHL-a反演模型,应用于无人机高光谱数据进行CHL-a的反演监测。结果表明:通过分析单波段、一阶微分值与CHL-a参数,发现使用674 nm、570nm波长构建的一阶微分与水体CHL-a构建的线性反演模型,具有较高的相关性与精度,可满足对小中型地表水体叶绿素的快速反演监测。基于反演结果构建空间分布图,实现水质参数的可视化。

基于时间序列的鄱阳湖Chl-a预测模型优化构建

为提升鄱阳湖富营养化评估与预测适应性,选择湖区代表性点位2012—2020年的逐月监测作为模型样本训练时间。模型自变量选择水体主要理化参数,并计算其中水质参数部分的综合水质标识指数(WQII)以评价鄱阳湖近年水质演变趋势。进一步地,分别建立响应变量为叶绿素a(Chl-a)的多元线性逐步回归方程(MLSR)、季节性自回归求和移动平均模型(SARIMA),并预测所有点位2020年6—8月份的浓度值,与实测值比较以检验两种模型适用程度。结果显示:各代表监测点的综合水质在2018年出现提升且在丰水期更优,平均WQII排序为康山(2.91)>都昌(3.01)>蚌湖(3.11)>蛇山(3.31);SARIMA模型相比于MLSR方程的Chl-a浓度预测准确度更高,为大型通江湖泊藻类爆发预警提供了优化理论参考。

基于ISSA-SVR算法的近海典型水质参数遥感反演

化学需氧量(COD)和叶绿素a (Chl-a)浓度作为与光谱相关的典型水质参数,是反映水体污染程度和富营养化程度的重要指标。支持向量回归模型(Support Vector Regression,SVR)适用于小样本,广泛用于近海典型水质参数的遥感反演,但也存在模型参数选择困难、容易陷入局部最优解的问题。针对这一问题,本文构建融合反向学习和模拟退火的改进麻雀算法(Improved Sparrow Search Algorithm,ISSA),通过改进麻雀算法对SVR模型的惩罚系数和核函数参数进行参数寻优,提出了一种改进的支持向量回归模型(ISSA-SVR)。通过该模型利用实测水面光谱与水质参数数据建立COD和Chl-a浓度反演模型。利用Sentinel-2卫星遥感光谱数据对模型的精度进行验证,得到各水质参数浓度的反演精度。采用ISSA算法优化SVR建立的COD浓度预测模型和Chl-a浓度预测模型的平均相对误差(MRE)分别为20.02%和30.17%。反演结果均优于其他模型(线性回归、SVR和SSA-SVR模型)。实验结果证实,ISSA-SVR算法是实现COD浓度和Chl-a浓度遥感反演的有效方法,可为我国近海典型水质参数遥感反演及后续水体科学管理提供参考。

利用Sentinel-2卫星影像反演骆马湖叶绿素a浓度

本文利用2023年6月15日采集的骆马湖水质参数、光谱数据和同期Sentinel-2影像数据,构建了基于数学统计和机器学习方法的水质反演模型,并对骆马湖叶绿素a(Chl-a)浓度进行了定量反演。通过对比分析,选定了性能最优的模型分析骆马湖Chl-a浓度状况。研究发现,Chl-a浓度与Sentinel-2影像B5和B9波段表现出较高的相关性,且经波段组合处理后,相关性进一步提升。在Chl-a反演模型中,FA-SVR模型相较于传统数学统计回归模型及其他机器学习模型(FA-RF、FA-XGBoost)表现出最高的精度(R^(2)=0.86,RMSE=2.77,MAE=2.10)。反演结果揭示,骆马湖东北部近岸区域的Chl-a浓度较高,这可能与北部水域存在鱼塘养殖、水体富营养高有关。本文突显了机器学习技术在提升水质遥感反演精度方面的重要应用价值,为湖泊水质监控和管理提供了重要的技术支撑。

象山港Chl-a的分布及其与环境因子关系研究

根据2012年5、8、10月在象山港海域的监测资料,分析了该海域Chl-a的时空分布特征,及其与水温、盐度、溶解氧、COD以及营养盐等主要环境因子的相关关系。结果表明：象山港海域Chl-a浓度在0.27~27.29μg/L范围内,平均浓度为2.80μg/L。海域Chl-a浓度的季节性变化非常明显,峰值出现在夏季,而秋季浓度最低;并且具有从内湾港顶向港外逐渐递减的空间分布规律。相关性分析显示,在夏季,影响Chl-a含量的主要因素是盐度和溶解氧;在秋季,Chl-a与活性磷酸盐和硅酸盐间的相关性非常显著,但与无机氮（DIN）不具显著意义相关关系。总体看来,象山港海域Chl-a的时空分布受到海域水动力条件、海水养殖活动等的共同影响,夏季还可能受到台风影响。

基于GCN和TPA混合模型的青草沙水库叶绿素a浓度预测方法

青草沙水库是上海重要饮用水水源地,面临较大藻华暴发风险,实现对水库叶绿素a(Chl-a)浓度的准确预测,对于饮用水安全保障至关重要。文章以青草沙水库为对象,提出了一种图卷积时间模式注意力网络混合模型(GC-TPA),首先利用时间模式注意力(TPA)机制学习水质数据的时间依赖性,其次使用图卷积网络(GCN)学习不同水质参数之间的关系。另外,为进一步提高模型的预测精度,引入完全自适应噪声集合经验模态分解(CEEMDAN)以降低模型的滞后性,同时使用多层感知机(MLP)学习Chl-a浓度的突变。结果表明:(1)引入的GCN模块显著增强了TPA对Chl-a的预测能力,结合CEEMDAN和MLP的帮助,模型性能进一步提升,以纳什效率系数作为评价指标,混合模型的24 h预测精度较单独TPA提升了56.5%;(2)与单独TPA和长短期记忆网络(LSTM)的对比试验表明,在更长的预测周期(48 h)上,GC-TPA虽然精度下降,但仍表现最好,48 h预测平均绝对误差和均方误差比LSTM低25.5%和24.0%,比TPA低4.92%和8.40%;(3)GCN模块与MLP模块在结果预测中发挥了不同的作用,GCN模块增强了TPA的特征学习能力,提高了模型对Chl-a浓度变化趋势的预测精度,而MLP则对Chl-a的突变较为敏感。研究所提出的GC-TPA混合模型在青草沙水库Chl-a浓度短期预测中表现良好,可为水库水质管理提供支撑。

BP神经网络在水库蓝藻水华预测中的应用

水体污染会造成水中营养物质过剩,藻类大量繁殖,从而加速水体富营养化。本文对水质自动监测数据的水温(T)、pH、溶解氧(DO)、电导率(W_cond)、浊度(Turb)、高锰酸盐指数(COD_(Mn))、氨氮(NH_(3)-N)、总磷(TP)、总氮(TN)9个因子进行分析,建立预测叶绿素a(Chl-a)浓度的BP神经网络预测模型,并对样本数据做主成分分析得出前5个主成分累计贡献率达83.59%,通过不同变量的网络模型筛选出pH、DO和TP是预测Chl-a浓度的3个主要影响因子。结果表明,模型预测值和实测值相关系数R2达到0.972,为Chl-a浓度预测提供了一种环保、安全、可靠的技术方法。

A field study on the conversion ratio of phytoplankton biomass carbon to chlorophyll-a in Jiaozhou Bay,China

A one-year field study was conducted to determine the conversion ratio of phytoplankton biomass carbon （Phyto-C） to chlorophyll-a （Chl-a） in Jiaozhou Bay, China. We measured suspended particulate organic carbon （POC） and phytoplankton Chl-a samples collected in surface water monthly from March 2005 to February 2006. The temporal and spatial variations of Chl-a and POC concentrations were observed in the bay. Based on the field measurements, a linear regression model Ⅱwas used to generate the conversion ratio of Phyto-C to Chl-a. In most cases, a good linear correlation was found between the observed POC and Chl-a concentrations, and the calculated conversion ratios ranged from 26 to 250 with a mean value of 56 μg ·μl^-1. The conversion ratio in the fall was higher than that in the winter and spring months, and had the lowest values in the summer. The ratios also exhibited spatial variations, generally with low values in the near shore regions and relatively high values in offshore waters. Our study suggests that temperature was likely to be the main factor influencing the observed seasonal variations of conversion ratios while nutrient supply and light penetration played important roles in controlling the spatial variations.

Estimation of chlorophyll-a concentration in the Zhujiang Estuary from SeaWiFS data

The chlorophyll-a concentration is generally overestimated for the southern China coastalwaters if the default algorithm of the SeaDAS is employed. An algorithm is developed for retrieval of chlorophyll-a concentration in the Zhujiang Estuary, Guangdong Province, China, by using simulated reflectance data. The simulated reflectance is calculated corresponding to the SeaWiFS wavelength bands, via a general model by inputting measured water components, i.e. , the suspended sediment, chlorophyll-a, and yellow substance (DOC) concentration data of 130 samples. Empirical relationships of the chlorophyll-a concentration to 240 different band combinations are investigated based on the simulated reflectance data, and the band combination, R_5R_5/R_3R_4, is found to be the optimum one for the development of an algorithm valid for the Zhujiang Estuary. This algorithm is then employed to determine the chlorophyll-a concentration from SeaWiFS data. The estimated concentrations have a better accuracy than those obtained from the SeaDAS default algorithm when compared with sea truth data. The new algorithm is demonstrated to work well and is used to derive a series of image maps of the chlorophyll-a concentration distribution for the Zhujiang Estuary and adjacent coastal areas.

Response of Sea Surface Temperature to Chlorophyll-a Concentration in the Tropical Pacific:Annual Mean,Seasonal Cycle,and Interannual Variability

The response of the upper-ocean temperatures and currents in the tropical Pacific to the spatial distribution of chlorophyll-a and its seasonal cycle is investigated using a coupled atmosphere-ocean model and a stand-alone oceanic general circulation model.The spatial distribution of chlorophyll-a significantly influences the mean state of models in the tropical Pacific.The annual mean SST in the eastern equatorial Pacific decreases accompanied by a shallow thermocline and stronger currents because of shallow penetration depth of solar radiation.Equatorial upwelling dominates the heat budget in that region.Atmosphere-ocean interaction processes can further amplify such changes. The seasonal cycle of chlorophyll-a can dramatically change ENSO period in the coupled model.After introducing the seasonal cycle of chlorophyll-a concentration,the peak of the power spectrum becomes broad,and longer periods（3 years） are found.These changes led to ENSO irregularities in the model. The increasing period is mainly due to the slow speed of Rossby waves,which are caused by the shallow mean thermocline in the northeastern Pacific.

Phytoplankton size class derived from phytoplankton absorption and chlorophyll-a concentrations in the northern South China Sea

A previously developed model was modified to derive three phytoplankton size classes (micro-, nano-, and pico-phytoplankton) from the overall chlorophyll-a concentration, assuming that each class has a specific absorption coefficient. The modified model performed well using in-situ data from the northern South China Sea, and the results were reliable and accurate. The relative errors of the size-fractioned chlorophyll-a concentration for each size class were: micro-:21%, nano-:41%, pico-:26%, and nano+pico:23%. The model was then applied on ocean color remote sensing data to examine the distribution and variation of phytoplankton size classes in northern South China Sea on a large scale.

Field Detection of Chlorophyll-a Concentration in the Sea Surface Layer by an Airborne Oceanographic Lidar

An airborne oceanographic lidar, with a frequency-tripled Q-switched Nd： YAG laser of 355 nm, has been designed to measure chlorophyll-a （Chl-a） concentration in the sea surface layer by the Ocean Remote Sensing Institute, OUC. The field experiment was carried out in the bay which is located south of the Liaodong Peninsula on the 10th of September 2005. After the flight, the raw data were processed and analyzed by the fluorescence-to-Raman ratio method with seawater attenuation coefficients calculated from signal profiles. The results of Chl-a concentration sea water were also compared with those of Chl-a concentration by measurements by lidar are shown. The measurements in clear a Moderate Resolution Imaging Spectroradiometer （MODIS）.

Application of Bayesian regularized BP neural network model for analysis of aquatic ecological data—A case study of chlorophyll-a prediction in Nanzui water area of Dongting Lake

Bayesian regularized BP neural network（BRBPNN） technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of the network were obtained. After the selection of input variables using stepwise/multiple linear regression method in SPSS i1.0 software, the BRBPNN model was established between chlorophyll-α and environmental parameters, biological parameters. The achieved optimal network structure was 3-11-1 with the correlation coefficients and the mean square errors for the training set and the test set as 0.999 and 0.000？8426, 0.981 and 0.0216 respectively. The sum of square weights between each input neuron and the hidden layer of optimal BRBPNN models of different structures indicated that the effect of individual input parameter on chlorophyll- α declined in the order of alga amount 〉 secchi disc depth（SD） 〉 electrical conductivity （EC）. Additionally, it also demonstrated that the contributions of these three factors were the maximal for the change of chlorophyll-α concentration, total phosphorus（TP） and total nitrogen（TN） were the minimal. All the results showed that BRBPNN model was capable of automated regularization parameter selection and thus it may ensure the excellent generation ability and robustness. Thus, this study laid the foundation for the application of BRBPNN model in the analysis of aquatic ecological data（chlorophyll-α prediction） and the explanation about the effective eutrophication treatment measures for Nanzui water area in Dongting Lake.

Tempo-spatial Variation of Nutrient and Chlorophyll-a Concentrations From Summer to Winter in the Zhangzi Island Area (Northern Yellow Sea)

Nutrient and Chlorophyll-a （Chl-a） concentrations were investigated monthly along three transects extending from a mariculturc area to open waters around the Zhangzi Island area from July to December 2009. The objective of this study is to illus- trate food availability to the bottom-sowed scallop Patinopecten yessoensis under the influences of the Yellow Sea Cold Water Mass （YSCWM）, freshwater input and feedbacks of cultivated scallops. Significant thermal stratification was present in open waters from July to October, and salinity decreased in July and August in surface layers in the mariculture area. Nutrient concentrations increased with depth in both areas in summer, but were similar through water column in November and December. On average, nutrient in- creased from summer to autumn in all components except ammonia. Nutrient concentrations lower than the minimum thresholds for phytoplankton growth were present only in upper layers in summer, but stoichiometric nitrogen limitation existed in the entire inves- tigation period. Column-averaged Chl-a concentration was lower in open waters than in mariculture area in all months. It increased significantly in mariculturc area in August and October, and was less variable in open waters. Our results show that nutrients limita- tion to phytoplankton growth is present mainly in upper layer in association with stratification caused by YSCWM in summer. Freshwater input and upwelling of nutrients accumulated in YSCWM can stimulate phytoplankton production in mariculture area. Farming activities may change stoichiometric nutrient ratios but have less influence on Chl-a concentration.

Relationships of Dissolved Oxygen with Chlorophyll-a and Phytoplankton Composition in Tilapia Ponds

This study investigated the relationships among the parameters of dissolved oxygen, chlorophyll-a and phytoplankton composition in tilapia ponds. Each pond (a total of 18 ponds) was sampled once in the dry, winter season between January and March and again early in the rainy season between May and June. The data were analyzed by examining correlations among parameters as affected by season, altitude and culture system. Observations were made at sites located in 5 selected provinces of northern Thailand: Chiangrai, Chiangmai, Phayao, Lampang and Nakornsawan. Mean elevation of these areas range from 25 to 582 meters above sea level (masl) and were categorized into low (400 masl) elevation sites. Ponds were 0.8 - 2.0 m deep, 0.16 - 0.64 ha in area and could be further categorized into high and low input systems.Mean air temperature in winter ranged between 16.5°C - 35.8°C while mean water temperature ranged between 25.5°C - 27.1°C. In rainy season, air temperature ranged between 22.0°C - 37.3°C and water temperature ranged between 29.4°C - 31.8°C. The amount of chlorophyll-a in both seasons were comparable (p > 0.05), but chlorophyll-a in high input system was significantly higher (p 0.05) than in low input ponds. Only weak correlation was found between chlorophyll-a, DOmax and DOmin. Multifactor-ANOVA was used to analyze the difference of total bacteria and filamentous cyanobacteria in ponds based upon elevation, culture systems and season. Result shows that there is a significant interaction observed between elevation, culture system and season (p 0.05). Species diversity and composition of phytoplankton in fish ponds in 2 seasons revealed the presence of 90 genera of phytoplankton under all 7 divisions. Divisions Chlorophyta and Cyanophyta had the most number of genera identified in both seasons with Pediastrum spp., and Scendesmus spp., and Anabaena spp. as dominant genera/genus, respectively.

Effect of Chlorophyll-a Spatial Distribution on Upper Ocean Temperature in the Central and Eastern Equatorial Pacific

Effect of the spatial distributions of chlorophyll-a concentration on upper ocean temperature and currents in the equatorial Pacific is investigated through a set of numerical experiments by using an ocean general circulation model. This study indicates that enhanced meridional gradient of chlorophyll-a between the equator and off-equatorial regions can strengthen zonal circulation and lead to a decrease in equatorial sea surface temperature （SST）. However, the circulation changes by themselves are not effective enough to affect SST in the equatorial cold tongue （CT） region. The comparison between the experiments indicates that the CT SST are more sensitive to chlorophyll-a distribution away from the equator. The off-equatorial chlorophyll-a traps more solar radiation in the mixed layer, therefore, the temperature in the thermoeline decreases. The cold water can then be transported to the equator by the meridional circulation within the mixed layer. Furthermore, the relation among CT SST, the surface heat flux, and the equatorial upwelling are discussed. The study implies the simulation biases of temperature on the equator are not only related to the local ocean dynamics but also related to some deficiency in simulating off-equatorial processes.

Study on the Relationship between Chlorophyll-a and Environmental Impact Factors in Deep Reservoir in Karst Areas——Taking Aha Reservoir for Example

[Objective] The aim was to study the relationship between chlorophyll-a and environmental impact factors in deep reservoir in Karst Areas.[Method] Taking Aha Reservoir for example,the changes of chlorophyll-a content and its relationship with environmental impact factors were researched,and the water quality of Aha Reservoir was assessed by means of modified Carlson trophic state index.[Result] Chlorophyll-a content in Aha Reservoir was higher in March,April,May and September and lower from June to August,and there was obvious seasonal variation,namely its variation trend was spring>summer>autumn>winter.In addition,chlorophyll-a content showed extremely significant correlation with transparency and dissolved oxygen (P<0.01) and significant correlation with ammonia nitrogen (P<0.05).Meanwhile,trophic state index was higher in whole year in Aha Reservoir which was in moderate trophic state in August,October and November and eutrophic state in other months.[Conclusion] The study could provide scientific reference for discussing the mechanism of lake eutrophication.

Spectral Geometric Triangle Properties of Chlorophyll-A Inversion in Taihu Lake Based on TM Data

The main objective of this study was to develop and validate the applicability of the Area Chlorophyll-a Concentration Retrieved Model (ACCRM), Height Chlorophyll-a Concentration Retrieved Model (HCCRM), Angle Chlorophyll-a Concentration Retrieved Model (AgCCRM), and Ratio Model of TM2/TM3 (RM) in estimating the chlorophyll-a concentration in Case II water bodies, such as Taihu Lake in Jiangsu Province, China. Water samples were collected from 23 stations on the 27th and 28th of October, 2003. The four empirical models were calibrated against the calibration dataset (samples from 19 stations) and validated using the validation dataset (samples from 4 stations). The regression analysis showed higher correlation coefficients for the ACCRM and the HCCRM than for the AgCCRM and the Ratio Model;and the HCCRM was slightly superior to the ACCRM. The performance of the ACCRM and the HCCRM was validated, and the ACCRM underestimated concentration values more than the HCCRM. The distribution of chlorophyll-a concentrations in Taihu Lake on October 27, 2003 was estimated based on the Landsat/TM data using the ACCRM and the HCCRM. Both models indicated higher chlorophyll-a concentrations in the east, north and center of the lake, but lower concentrations in the south. The accuracy of results obtained from the HCCRM and the ACCRM were also supported by the validation dataset. The study revealed that the HCCRM and the ACCRM had the best potential for accurately assessing the chlorophyll-a concentration in the highly turbid water bodies.

The algorithms of chlorophyll-a concentration for CMODIS

The algorithms of extracting chlorophyll-a （Chl-a） concentration have been established for Chinese moderate resolution imaging spectrometer （CMODIS） mounted on Shenzhou-3 spaceship launched on 25 March 2002. The CMODIS is an ocean color sensor with 30 visible channels and 4 infrared channels, much different from other ocean color satellites and needs new algorithms to process data. Three models of Chl-a concentration were established based on Chl-a data retrieved from sea-viewing wide field-ofview sensor （SeaWiFS）, with the average relative errors of 26. 6%, 24%. 0% and 33.5%, respectively. This practical and economic approach can be used for developing the algorithms of Chinese ocean color and temperature sensor （COCTS） on the satellite Haiyang-1 to derive the Chl-a concentration concentration distribution. The applicability of the algorithms was analyzed using some in situ measurements. Suspended sediment is the main factor influencing the accuracy of the spectral ratio algorithms of Chl-a concentration. The algorithms are suitable to using in the regions where suspended sediment concentrations （ SSC ） are less than 5 g/m^3 under the condition of relative error of Chl-a concentration retrieval within 35%. High concentration of suspended sediment leads to the overestimate remote sensing retrieval of concentration of Chl-a, while low-middle SSCs lead to the low Chl-a concentration values using the spectral ratio algorithms. Since the accuracy of Chl-a concentration by the spectral ratio algorithms is limited to waters of Case 2, it is necessary to develop semi-analytical models to improve the performance of satellite ocean color remote sensing in turbid coastal waters.

Numerical Study of the Three Gorges Dam Influences on Chlorophyll-a in the Changjiang Estuary and the Adjacent East China Sea

The impoundment of the Three Gorges Dam(TGD)has altered downstream hydrological characteristics and sediment discharge,and it has caused ecological impacts,such as changes in chlorophyll-a(Chl-a)in estuaries and coastal oceans.To investigate the TGD's influence on Chl-a's interannual and seasonal variations in the Changjiang Estuary and the adjacent coastal East China Sea,a physical-biogeochemical model was developed with numerical experiments covering a decade,including TGD's preperiod(pre-TGD,1999–2003)and post-period(post-TGD,2004–2008).The modeling results demonstrate an annual increase in the regional average Chl-a from pre-to post-TGD,with the largest increase reaching up to 20.8%in spring in the nearshore region beyond the Changjiang mouth.The spatial variations in Chl-a are high,with the largest variation being observed around the 20–40 m isobaths.The increase in Chl-a is influenced by improved light and nutrient conditions,which is attributed to dam construction and fertilization by human activities.A decline in nitrogen-phosphorus fertilizer usage ratio along the Changjiang River watershed after the TGD's impoundment is another factor that influences the Chl-a's variation.The modeling results also show severe NO3 and PO4 imbalances with a larger N/P ratio during the post-TGD period,especially in regions with large Chl-a increases.Moreover,the greater increase in the usage of phosphorus fertilizer than nitrogen fertilizer influences Chl-a's variation and has potential promotion effects on the outbreak of harmful algal bloom events.