Using Digital Repeat Photography And MODIS Satellite Remote Sensing To Monitor Phenological Seasonal Changes Of Sichuan Forests And Establish Its Management System

Phenology is a comprehensive indicator of climatic and environmental change of the most intuitive, sensitive and diagnosis of "fingerprint", plant phenological information was not only reflect local environmental conditions, but also reflects the accumulation of environmental conditions over a period of time. With the development of science and technology, research on plant phenology is no longer confined to the traditional means, satellite remote sensing and digital camera were used in the study plant phenology. Combined with satellite remote sensing technology, digital repeat photography using digital camera technology can save the cost of labor and monitoring, but also study changes of plant phenology in a large area.This study selects the forest of Sichuan province and the typical autumn leaf specices Ginkgo, Metasequoia in Dujiangyan Lingyanshan Forest Park. Based on self-designed phenological data acquisition system, this system can accomplish the automatically shoot digital images, automatically acquisition. Combing with the MODIS surface reflectance products, the paper constructs a study of forest phenology of Sichuan. Near ground digital camera time-series data and MODIS satellite remote sensing time series images were analyzed using ENVI/IDL, ArcGIS, TimeSAT software, analys the possibility of monitoring forest phenology at landscape and regional scales with these two plant phenological monitoring means. From these researches, main results are drawn as follows:(1)Using the software that can automatically capture digital repeat photography to get visible and infrared photographs of ginkgo, Metasequoia. Computing ExG, Gcc, NDVI and EVI Etc., a total of seven kinds of vegetation indices of the digital photos, shows only ExG, Gcc, NDVI can accurately reflect the growth trends of forest vegetation during the year from each vegetation index time series curve.(2) Smoothing the vegetation index with double logistic model, obtained a more reasonable vegetation index time-series data. Using dynamic threshold method to extract the phenological critical:the start of vegetation growing season (SOS), the end of vegetation growing season (EOS), growing season length (LOS) and the DOY of the maximum growth (MOE). The results showed that the phenological critical based on ExG and Gcc are closer:SOS is about 130 days; EOS s about 280 days; LOS is about 150 days; MOE is about 180 days. The result shows that it is feasible based on a digital camera to monitor vegetation canopy phenology changes, and the vegetation index based on visible band is better than infrared band at the characterization of climate change information.(3)By preprocessing MODIS 8 day surface reflectance product data, and calculating NDVI, EVI, ExG and Gcc vegetation index, the vegetation index time-series curve shows that NDVI, EVI, ExG and Gcc were able to correctly reflect the growth trends of forest vegetation during the year.(4)Smoothing vegetation indices EVI with different methods, found the S-G filtering method is superior to other methods. Using dynamic threshold methodto extract the phenological critical of Sichuan forest, compared to four vegetation indices, found that the EVI is superior to the other three vegetation indices in reflecting the changes of forest vegetation phenology.(5)Using the same method extracts satellite-based remote sensing of SOS, EOS, LOS and MOE. According to the spatial distribution of the phenology critical period of Sichuan forest in 2014. Overall, the spatial distribution in Sichuan forest phenology of rendering the following rules:SOS:interior basin, the northern margin of the basin (90th to110th)< western basin, the southern basin, and basin southwest (110th to 130th)< subalpine, northwest (150th to 170th); EOS:basin in the southwest (310th to 330th)> interior basin, north margin of the basin (290th to 310th)> western basin, south of the basin (270th to 290th)> subalpine, northwest (250th to 290th); LOS interior basin, north of the basin (190th to 210th)> western basin, southern basin, southwest basin(150th to 170th)> subalpine, northwest (110th to 130th); MOE:in addition to the southwestern margin of the basin there are more than 30% of the forest area of more than 220 days, the other most of the forest concentrate on the day 180-200.(6)Research on the relationship between altitude and phenology critical period, the results show that below the 2000m, for every increase of 200m, SOS delayed 3.1 days, EOS delayed 5.1 days, LOS extended 2 days, MOE delayed 0.8 days; from the 2000m to the 4000m, for every increase of 200m, SOS 3.4 days in advance, EOS 8.2 days in advance, LOS 4.8 days shorter, MOE shortened 1.5 days. above 4000m, forest vegetation phenology changes with altitude law obvious.(7)Comparing the vegetation phenology critical period derived form two different monitoring tools, found that the results are different because of the different methods and different vegetation indices. Overall, the EVI based on satellite-based remote sensing and the ExG and Gcc based on the digital cameras are more suitable for monitoring forest vegetation phenology.(8)Using IDL、C# develops the forest phenological remote sensing monitoring system. This system includes a digital camera to monitor automated management software and forest phenology of remote sensing monitoring system, the main achievement of digital photos automatically capture, display vector data and image data, basic map manipulation tools, basic image processing, extracting digital photos channel value, MODIS data preprocessing, vegetation phenology parameter extraction and network online monitoring.