Zoning of high standard farmland construction based on local indicators of spatial association

Abstract: How to define or partition the actual construction areas is an important research subject. Zoning of high standard farmland construction is not merely a process of quantitative analysis; the natural, economic, social and spatial attributes of farmland should all should be taken into account. In this paper, we tried to improve the rationality of zoning for high standard farmland construction by the spatial structural characteristics of farmland, and by exploring the new zoning method which designated the construction area through spatial analysis. Taking Guangning county in Guangdong Province as a case study, the whole process of the study was composed of three steps: 1) evaluating the quality of farmland comprehensively at map-spot level, 2) taking the result of evaluation at map-spot level as the spatial variation, and exploring the spatial distribution attribute in scale of village through the introducing of Local Indicators of spatial Association, and 3) partition the construction areas according to the results of LISA. The results show that the overall average level of farmland quality in Guangning was relatively high, the high score of map-spots mainly located in the southeast of county, while the low score map-spots spread at the north and southwest. Four classes can be distinguished for better comparable recognizing, and the quality of four classes which contained the first-order, second-order and third-order were about 4366.01, 14796.25 and 3352.96 hm2, respectively. The first-order was the best and the third-order was the lowest comparatively. Moreover, a relatively strong spatial autocorrelation existed among the distribution of farmland quality in Guangning in scale of village, and the positive autocorrelation was superior than passive autocorrelation. In detail, 75.96% of villages had positive autocorrelation which contained the High-High type and Low-Low type(the High-High type meant the village with higher farmland quality was surrounded by villages with same higher farmland quality, in contrast, the Low-Low type indicated the villages with lower farmland quality located together), and 24.04% of them appeared passive autocorrelation which contained the High-Low type and Low-High type(the High-Low type signified the villages with lower farmland quality surround the village with higher quality, and on the contrary the Low-High type meant the villages with higher farmland quality surrounded the village with lower quality). The distribution of spatial autocorrelation types in village level was basically identical with the map-spot of farmland. About 86.34% of first-order farmland shared the same spatial location with the High-High type, while 51.54% of third-order farmland located in the distribution of Low-Low type. In addition, a rational zoning scheme arose by splitting the county into four regions, which comprised of the optimization region, key region, reserve region and general region. These were consistent with the results of spatial autocorrelation. The optimization region mainly located in "Hengshan-Nanjie-Paisha" groups, the key region mainly consisted of the "Binheng-Hengshan", "Luogang-Jiangtun" and "Shizui-Muge" groups, the reserve region approximately located around the key-region distribution, while the "Beishi-Chikeng-Kengkou-Gushui" group was the typical spreading area for general region. Various kinds of improved and reclaimed measures for farmland protection can be adopted according to different regions or diverse types of spatial autocorrelation. Based on the results showed above, conclusions included that the local indicator of spatial association was an effective method and can reveal the agglomerate features of farmland. On the other hand, the zoning scheme reflected the three-in-one protecting conception by taking the spatial agglomeration into account. The quantity, quality and agglomerate attributes of farmland can be combined together preferably. The paper introduced a new perspective for farmland protection and reclamation.