Abstract:Mangrove ecosystems are important for socioeconomic development and environmental protection but are experiencing sharp declines around the world. Asia has the most mangroves in area around the world but also has the largest human populations that threaten the ecosystem. Where mangrove is distributed and how its range changes are important topics for research throughout the world. Although many efforts have been made to study mangrove, more detailed information is still needed for better managing and protecting mangrove ecosystems. To analyze the mechanism of spatial dynamics of mangrove in Guangxi Beibu Gulf, China, from 1960 to 2010, multi-source remote sensing data were used to extract spatial information on mangroves in various years. The spatial distribution characteristics were analyzed and the spatial dynamic mechanisms of mangroves from 1960/1976 to 2010 were analyzed with a patch-based method. The results indicated that there were 1020, 829, 1094, 1718, and 1712 patches of mangrove in 1960/1976, 1990s, 2001, 2007, and 2010, respectively, and the areas were 9062.5, 7430.1, 7015.4, 6743.2, and 7054.3 hm2 in these 5 years, respectively. The region lost 22.16% of its mangrove area from 1960/1976 to 2010, with an annual loss rate of 0.53%. In the meantime, the patch-number increased 67.8%; the average patch-area decreased from 8.9 hm2 in 1960/1976 to 4.1 hm2 in 2010, and the number of big patches became smaller over time. Only 24 of 1020 patches (2.4%) in 1960/1976 remained unchanged till 2010, almost all the patches had changed to a certain degree. The results also indicated that lost (46.1%), fragmented (40.4%), and shrunken (13.5%) mangroves were the major components of the loss in mangrove area, whereas new (70.0%) and fragmented (29.9%) patches of mangrove were the main reasons for patch-number increases from 1960/1976 to 2010. The dynamics of patch-number and area changes varied both spatially and temporally. Shrimp and salt pond construction (80.0%), infrastructure construction (10.9%), and agriculture (9.1%) were the major causes of mangrove loss, and incensement of patch-number was due to natural processes (92.6%) such as extreme weather, flooding, sediment deposition, and growth/succession and plantations (7.4%) The causes of lost area varied both spatially and temporally. Changes in patch-number were mainly due to the natural processes and procedures of new patch (39.6%) and disappeared (-9.1%), and shrimp and salt pond construction and procedures of disappeared (-15.3%) and fragmented (14.5%); changes in mangrove areas were caused mainly by natural processes acted on the new (17.5%), expanded (12.6%), and disappeared (-6.1%) patches; shrimp and salt pond construction contributed to the disappeared (-14.8%), fragmented (-13.9%), and shrunken (-6.6%) mangrove patches. This study demonstrates the effectiveness of methodology that integrates natural processes and human activities into analysis of mangrove spatial distribution and its changes over time, and such a method can be used for in-depth mangrove research in other regions of the world.