Abstract Hydrologic models of wetlands enable water resources managers to quantify the environmental and societal roles of wetlands and manage them in ways that sustain their valuable services. However, reliable wetland models require data that are not typically available from in-situ measurements. In this article, we use satellite information from MODIS (500-meter, 8-day composite land surface reflectance product) and limited ground data to quantify the seasonal and inter-annual changes of wetland extent. This information is used to calibrate a new, non-parametric land cover classification approach. Extensive tests demonstrate that the new approach performs well in (i) classifying accurately land cover classes and (ii) delineating reliably seasonal and inter-annual wetland area changes. The new approach is applied to the Sudd wetland in South Sudan, a vast wetland of vital socioeconomic and environmental services, toward developing better, policy-relevant information and tools.

中文翻译：

---

使用 MODIS 进行土地覆盖分类和湿地淹没制图

摘要 湿地水文模型使水资源管理者能够量化湿地的环境和社会作用，并以维持其宝贵服务的方式对其进行管理。然而，可靠的湿地模型需要通常无法从现场测量中获得的数据。在本文中，我们使用 MODIS 的卫星信息（500 米，8 天复合地表反射率产品）和有限的地面数据来量化湿地范围的季节和年际变化。此信息用于校准新的非参数土地覆盖分类方法。广泛的测试表明，新方法在（i）准确分类土地覆盖类别和（ii）可靠地描绘季节性和年际湿地面积变化方面表现良好。