| Tidal flat possessing biological resources, energy and so on plays a significant role for human life and social economy. Intertidal resources are currently used for marine aquaculture, reclamation, port construction and ocean energy development. Human activities may change the environment of coastal waters. In order to evaluate the influences, we need to use effective means to monitor the coastal tidal-flat terrain and wetland environment changes. However, water content of intertidal surface changes severely in tidal flat. It affects the accuracy of sediment information extraction by means of remote sensing. Therefore we need an effective method to extract water content information of tidal flat. Taking two pieces of intertidal flat as experimental sample source, which are located at south of dafeng and xinyang port in Jiangsu province. We unified moisture spectral acquisition environment and standardized key procedures of operating. And then, we constructed5remote sensing inversion models based on the relationship between moisture content and spectral reflectance. At last, through analyzing the advantages and disadvantages of above models, we obtained the tidal flat moisture content spatial distribution. The main contents of this thesis are as follows:(1) The inner wall of culture dish which is exposed in air should be painted with black color for reducing light pollution from the wall of culture dish, otherwise reflectance of sediment would approximately increase4.1%; Sediment spectra were sensitive to the change of observation height under relatively small observation IFOV, however, the spectra measured under different observation height possessed significant linear correlation while the correlation coefficient reached0.954; The placement of culture dish effected reflectance of dry sediment diminutively, on the contrary, reflectance had an obvious change when culture dish moved to artificial light source, in the latter condition, spectra of sediment measured from multiple nadir location should be averaged to reduce the effects of sediment surface heterogeneity on optical measurements.(2) The spectral characteristics of the sediments demonstrated that under dry soil conditions the reflectance of different types are different, but the overall curves are similar and parallel. The reflectance increases simultaneously with wavelengths increasing in the region of427nm-1790nm and2002nm-2153nm, and has different trend in2133nm-2335nm region. In wet soil conditions, the reflectance of sediments decreases gradually with the increase of water content. Typically, at lower water content conditions, long wavebands (such as1679nm and2034nm band) are most sensitive to the changes of water content in sediments. When moisture content is higher, in the933nm～1346nm range, wavelengths are more sensitive to changes. It is able to provide a reference to construct water content prediction models.(3) The single band model showed that except sandy silt model, the R2is0.543, remaining sediment water content models are all larger than0.62,so single band inversion model have little effect on inversion results.(4) Band combination model demonstrated that NDWCI index model is very sensitive to the noise of sediment background (R2=0.675), so its predictive capacity is the worst. RWCI index can enhance the difference between water and sediments, but when the moisture content in lower conditions, its distinguished ability gets significant reduce. SAWCI and EWCI index model introduced a parameter L, which can be used to adjust sediment background. According to model structure, we can define L to adjust new models possessing high predictive accuracy. The result showed that both index models can reduce the effect of sediment background. EWCI index model has better moisture estimates capacity, the sediment average value R2is0.818, especially when we ignored sediment type its R2reaches0.706and RMSE is0.059.(5) Standardizing the reflectance of different sediment types by using dry sediment reflectance to build moisture inversion models. It has been proved that this way is able to improve the predictive accuracy of water content and effectively reduce the dependency on sediment type. Building NEWCI index model to predict water content when the moisture content is0and0.3～0.6, we obtained higher precision than EWCI index model is0.032, and when the water content up to0.3, EWCI index much better than NEWCI. Taking into account the actual moisture content in tidal flat, we consider that NEWCI index model is more suitable to predict the sediment water content of tidal flat.(6) Due to the gentle slope to water content of sediments in the study area has little impact, and from east to west, particle size becomes fine gradually. It can be inferred that sediment components distribution is a major factor affecting the sediment water content. Contrasting predicting moisture layers of intertidal surface, derived from EWCI and NEWCI index models. NEWCI index model is more accurate to reflect tide groove high moisture area of moisture situation and NEWCI index model gets of research district moisture distribution situation and research district deposition property component distribution situation match. Therefore the moisture distribution derived from NEWCI index model is more accurate and can reflect the real situation of intertidal flat. |
PDF Full Text Request