| Multiple Cropping Index(MCI) is the main parameter to reflect the behavior of multiple cropping. And it is one of the important indicators to evaluate land use degree and efficiency at the time scale. Multiple cropping can increase the food production and plays an important role of ensuring the food security. Remote sensing techniques can monitor the change of the time-series vegetation index and reflect land crop conditions quickly and objectively at the large scale. Thus, it may monitor the MCI changes. Using the remote sensing techniques to monitor the MCI can avoid the over-reliance on statistical data than before and reduce the errors such as statistical standards, data quality and human factors and so on. It provides the new direction to study the MCI.The study uses IDL programming language to implement wavelet transform denoising SPOT/VGT time-series NDVI dataset of the HuangHuaiHai Plain in 2007. It defines the multiple cropping and develops the method to distinguish the peak value combining the region farming calendar with statistical data. Finally, the study calculates the MCI of the five provinces (municipalities) in the HuangHuaiHai Plain. The result shows that it is one of the fast and effective means to monitor the MCI changes using the wavelet transform to denoise the time-series vegetation index and has higher reliability and operability. The main conclusion includes:(1) Combining the crop farming calendar with crop cultivated area derived from the statistics data, there are mainly single cropping a year and double cropping a year in the HuangHuaiHai Plain which the main crops include winter wheat, summer maize, cotton and peanut. It is observed that the cyclical change of the time-series NDVI corresponding to different crop growth stages which have a close correlation with multiple cropping. That is, a peak value indicates the single cropping a year and there should be two peaks where it is the double cropping a year.(2) According to the analysis of the experimental data, the difference of the peak values of the time-series NDVI dataset generally not less than 40% of the maximum value where there is double cropping a year, and the peak value would be greater than 140 for the single cropping a year. Therefore, attaching crop farming calendar and the dynamic threshold based on the second difference method, the time of the reasonable peak value should be late of April to early of October and the smaller peak value should be more than 60% of the maximum peak. For the single peak, the peak value ought to be more than 140.(3) It is selected that the COIFLET5 is the mother wavelet to denoise the time-series NDVI dataset based on a series of experiment of 30 agricultural remote sensing data in Hebei, Henan and Shandong provinces. Comparing the study result with statistics data and other similar studies results, there have the good agreements which are 0.97(comparing with five-points weighted average method),0.957(comparing with decision trees method),0.938(comparing with harmonic analysis of time series method), one of which the absolute error is the smallest between the wavelet transform method and decision tree method to monitor MCI. (4) The area of the double cropping a year is bigger than that of the single cropping a year in the HuangHuaiHai Plain. There have distinct differences in the spatial distribution map of the MCI. In the central and southern of the HuangHuaiHai Plain, cropping system is mainly the double cropping a year. In the north of the HuangHuaiHai Plain, as latitude increases, the temperature and precipitation cannot satisfy the needs of the double cropping a year and mainly maintain the single cropping a year. Among of the five provinces (municipalities), the MCI of Henan is the biggest which is 187.9, the cropping system of the spatial differences of Hebei and Tianjin is very significant, the cropping system is more complicated in Shandong, and there is mainly single cropping a year in the Beijing region. |
PDF Full Text Request