The Relationships Separately Between The Netprimary Productivity, The Evaportranspiration And The Change Of Landscape Pattern At Yellow River Delta

Net primary production (NPP) and evaportranspiration (ET) are both key components of the studies of global change. The great development and application of remote sensing promote the landscape pattern study and estimation methods of NPP and ET, both of which have developed many study methods and techniques and models. There is close relationship between landscape pattern and ecological progresses. Based on that principle is there also a relationship between landscape pattern changes and NPP and ET which stands for two ecological progresses? And if there is, how is it?This study chose Yellow River Delta (YRD) which is influenced greatly by human beings as the study area, first carried out the classifications of landscape types, then estimated NPP with the modified CASA model and ET with the SEBAL model, and explored the scale effects of ET and NPP with landscape ecological methods and geostatictics, finally analyze the statistic relationships between ET and NPP and landscape pattern indices such as NP, so that the impacts from landscape pattern to ET and NPP can be quantitatively analyzed.Main contents and results are listed as follows:Classification and analysis of the landscape pattern. The classifications were carried out based on Landsat TM remote sensing data in summer in 1992,1999 and 2008. Finally the YRD was classified into 15 kinds of landscape types, which are building zone, bare land, woods, grass, arable land, paddy field, lake & pond, saltpan, cultivation pond, river, tidal flat, salting, reed, swamp and meadow. Pattern analysis showed that the fragmentation degree decreased and the distribution of landscape types tend to more even in the past 16 years.Estimated ET with building SEBAL model and NPP with modified CASA model. Compared with the other studies, the results were credible. And as is the distribution trend of ET on each kind of landscape type:waterbeds like lake& pond, salt plan, cultivation pond and river hold the highest ET, then vegetations like woods, grass, reed, meadow and swamp, and building zone and bare land the lowest. While NPP: woods the highest, then the arable land, paddy field and other vegetations lower, and non-veg zones much lower, water the lowest.Explored the effective scales of ET and NPP with auto-correlation coefficient and semi-variance analysis. Then calculated pattern indices with moving window which side is right the effective scale. The result showed that ET and NPP showed obvious scale effect within the scale of 2,000m while none on the landscape scale.Analyzed the significant of relation ships between pattern indices and ET and NPP with partial-correlation analysis, and then built models with the step regression. It was found that NDVI and land surface temperature Ts are the main factors which separately impact ET and NPP, while MPFD and IJI also have significant correlations with ET and NPP.