Study On Spectral Model Of Plants Stressed By Hydrocarbon Microseepage

Hyperspectral remote sensing has been widely used in the hydrocarbonmicro-seepage information extraction, particularly in the low vegetation covered orbare areas. But it has barely been applied in vegetation covered area, vegetation hasbeen seen as interference information in existing applications. However, under theinfluence of underground oil and gas resources, the vegetation spectra will change. Atthe same time, vegetation spectra simulation by PROSPECT model has been widelyused, which provides a new technology to establish a coercion spectral model invegetation-covered area for detecting micro leaks information.Based on the principle of hydrocarbon microseepage, we simulated hydrocarbonmicroseepage environment in laboratory and foster wheat and cole.10differentvolume of gasoline were set to influence the vegetations respectively. In order tosimulate vegetation spectra by PROSPECT model, we acquired four different periodsof wheat and cole spectra and biochemical parameters regularly.By analyzing the changes of vegetation chlorophyll content, water content anddry matter content impact on oil and gas, the relationship between changes inbiochemical parameters of content and different gasoline volume was found not as asimple linear relationship. By contrast spectral angle difference between the controlgroup and stress vegetation spectra, we found that71%vegetation spectra impactingon oil and gas have taken place changes obviously. So we used classical physicalmodel-PROSPECT model to simulate spectral reflectance of the folium, analyzedthe impact of oil and gas on vegetation spectra from the perspective of the physicalmeaning. PROSPECT model, including input chlorophyll content, moisture content,dry matter content and structure parameter N of as the biochemical parameters. Weobtained the optimum structure parameters from traversing simulated spectra with themeasured spectral with the minimum error by setting the range and step size forstructure parameter. The optimal value of N of wheat is1.4-2.2, the optimum N valueof rapeseed is1-1.8, and then get the analog spectra under influence of differentvolumes of gasoline through PROSPECT model.The effect of different parameters on the analog spectra changes were analyzedby changing four parameters respectively, and the bands with large sensitivity weredetermined as sensitive bands for the influence of oil and gas. In order to establishstressed spectral model,9spectral characteristic spectral parameters in the sensitivewavelength range were analyzed, including the absorption position, the absorptionpeak width, red valley absorption depth, absorption area, symmetry, the first derivative maximum of red side, the height of green peak, red and green valley peakabsorption depth height ratio of the reflected peaks and valleys of red and green frontabsorption depth reflection peak height of a normalized value. The largest differentspectral characteristic parameters were obtained from stressed spectra and controlgroup. The highest correlation coefficient of the regression equation was obtained byregression analysis between gasoline volume and different characteristic parameters.The hydrocarbon microseepage vegetation stressed spectral model was established bythe characteristic parameter.Qingyang in Gansu Province was selected as study area. The area was depositedin a thickness about500-1000m in the Cambrian and Ordovician marine strata, thecontinental deposits was accepted in late Paleozoic, sedimentary was not accepted inthe late Proterozoic and early Paleozoic. The main structural feature is thedevelopment of the irregular nose structure but fault is not developed. Due to theuneven lifting platform base in southern basin development movement continuedlong-term depression, it developed to a large area of gentle inland lake until the lateTriassic. The lake was dying and the land was eroded by uplifting in late Triassic. Itbegan to sink again and developed of wide cover type fluvial flood plain depositionafter entering the Jurassic. Thus the two major oil strata were formed in Mesozoic,more than20oil fields has been found. On October14and15,2010, the CASI/SASIairborne hyperspectral data were acquired in the study area. The establishment ofhydrocarbon micro-seepage vegetation stress spectral model was applied on CASI/SASI data to extract the oil and gas microseepage information. Compared with theknown of oil and gas points, the micro-seepage information was identical with oilpoints. It proved the applicability and superiority of hydrocarbon microseepagestressed spectral model in the vegetation covered area for extracting microseepageinformation.