Yuping Lei

Estimating regional agricultural water use based on remote sensing data: a case study at Luancheng County of North China Plain

Sustainable management of water resources requires reliable information on regional evapotranspiration (ET) distribution, which is the largest output component of the hydrological cycle in North China Plain (NCP). In this work, we integrate a popular remote sensing technique with ArcGIS to build a ArcMap tool bar, named rGIS-ET, for estimating regional ET from Landsat TM/ETM+ data. The development of rGIS-ET enables quick processing of large amount of remote sensing and other spatial data. It also provides user-friendly interfaces for modeling, output display and result analyses. We use daily ET measurements from a weighting lysimeter in our experimental station to verify the performance of rGIS-ET. The verification confirms the reliability of ET calculation, whose errors during crop growing season are less than 10 %. We apply rGIS-ET to Luancheng County, a typical agricultural region in NCP, to demonstrate its utility for calculating regional ET and estimating agriculture water needs and ground water usage, both of which are critical to the design of an effective water resources management program for achieving sustainable development.

Calculation of albedo on complex terrain using MODIS data: a case study in Taihang Mountain of China

Albedo is one of the import parameters for the routine applications of various remote sensing models, and most methods for estimating albedo are only applicable in plain areas with relatively flat topography. Due to the existence of the undulant terrain in mountain areas, the incoming radiance can be decomposed into beam, diffuse, and terrain radiance, which are affected by atmospheric transmittance, solar zenith and azimuth angles, terrain slope and aspect, etc. The existence of partial shadow in low-resolution pixel makes the calculation of albedo more complicated in the mountain areas. This paper presents a rapid and operational method for estimating albedo on complex terrain using high-resolution Digital Elevation Model (DEM) and MODerate-resolution Imaging Spectroradiometer (MODIS) data with sensor view angles less than 20°. Shadow coverage ratio of the low-resolution map was retrieved from high-resolution shadow map, and was used for the correction of beam radiance. The parameters used for the atmospheric correction of narrowband beam radiance and weighting coefficients for the calculation of broadband albedo using MODIS with and without band 5 were developed based on the SMARTS2 simulation under different atmospheric conditions. Taken the Taihang Mountain in the North China as the study area, the calculated reflectance and albedo were compared with the MODIS reflectance and albedo products. The maximum deviations between estimated and blue-sky albedos were −0.0155 and +0.0151 in plain areas, and −0.00667 and 0.01747 in the mountain areas, respectively. Good agreements indicate that the procedure presented in this paper is applicable for the albedo calculation in the mountain areas using MODIS images and high-resolution DEM.

Monitored landscape change of Lake Baiyangdian wetland with dynamic reed area based on remote sensing

Lake Baiyangdian, a largest wetland ecosystem in North China Plain, has dried up on seven occasions since the 1960s. In recent years, more than one billion of cubic meters of water from upstream reservoirs and Yellow river have been transported to the lake to rescue the shrinking wetlands. Since the Lake Baiyangdian was actually composed of 143 small lakes and more than 70 villages with large or small area of cropland, dynamic distribution of aquatic plants in wetland such as reed and associated growth condition of these allowed to monitor the changes of wetland landscape and water quality to support the policy applications of water conveyance and wetland environmental treatment and control. Assisted with ground survey analyses and Landsat TM image, the MODIS 250 m time series Normalized Difference Vegetation Index (NDVI), given its combination of medium spatial and high temporal resolution, were applied to detect the unique rapid growth stage of reed in the spring from adjacent crops such as winter wheat, cotton, and spring maize, of which has a similar phenology in development of leaf area index, and dynamic reed areas were mapped in recent decade. Landscape changes of the wetland were analyzed using maps of reed area and hydrological data.

Evaluating crop land productivity using MODIS derived time serious field greenness and water index in North China Plain

Mapping grain crop land productivity that associated soil quality and crop field management are needed over intensively cropped regions such as the North China Plain to support science and policy application focused on understanding the current and potential capacity of regional food support. In this study, the crop growth dynamic presenting by time series field Greenness derived from MODIS 250 m data and soil moisture condition assessing by Normalized Difference Water Index (NDWI) derived by MODIS 250 m and 500 m data were combined to detect the temporal and spatial variability of productivity of winter wheat-summer maize field in the period 2000 to 2008 in Hebei and Shandong Province in North China Plain. Annual average NDVI levels, average levels of nine years and coefficients of variation of levels in the main growing season indicated corresponding crop growth condition and clearly presented spatial distribution of crop growth. Both the levels of NDWI and the coefficients of variation of the levels have almost same pattern of spatial distribution and correlations between two indexes levels were very high. The results of analysis of levels and coefficients of variation of levels of NDVI and NDWI shows the combination analysis of two indexes can be used to assess the levels of land productivity with a high spatial or temporal resolution .

RS-CGM: a spatial crop growth model based on GIS and RS

Spatial variability of crop growth often needs to be evaluated due to different soil conditions, weather patterns and crop information in a region. To simulate crop growth and productivity at a regional scale, a RS- and GIS-based crop growth model named RS-CGM was developed. The model calculates crop distribution, leaf area index, soil water content using remote sensing data that were integrated in crop growth module by inputting direct forcing variables, re-calibrating specific parameters, and correcting yield prediction using simulation-observation difference of a variable. The main RS-CGM components were intensively calibrated and verified against comprehensive field measurements of soil conditions, irrigation, evapotranspiration (ET), crop leaf area index (LAI) and yields. .The RS-CGM was applied to a county in the North China Plain to simulate winter-wheat yields in spatial and temporal dimensions. The model divides the simulating area into a number of crop growth elements and calculates each element with a set of parameters, then achieves the spatial crop yields and other concerned results aggregating to administrative regions. The simulated results show that the model can effectively express the spatial variety of yields in a region. And suggest that it was feasible to develop a spatial crop growth model combined with GIS, RS, and physiological process-oriented.

Integrated remote sensing and hydrological models for water balance in mountain watersheds

The evapotranspiration (ET) is one of the most important components of the water cycle in semi-arid Taihang Mountain region of North China. Due to significant changes in topography, the ET of this semi-arid region tends to vary dramatically both in time and space, which renders the accurate estimation of yearly or seasonal ET a difficult task. In current study, based on rGIS-ET v1.0, a regional ET model, by adding module of adjusting surface temperature in terrain, solar radiance terrain correction, and shaded relief, we improved the rGIS-ET a remote sensing model on ArcGIS platform for mapping ET distribution in such a semi-arid mountain area. With DEM of 30 meter and climate data, we run the model to estimate daily ET in mountain area using Landsat data. The ET distribution pattern estimated from Landsat data by rGIS-ET v2.0, was integrated into SWAT method to calculated daily ET at a high spatial resolution in the sub-watersheds. With input of the daily ET, SWAT simulated the annual flow of the component of water balance at sub-watersheds from 1995 to 2003. The variations of flows of annual ET were significant amount the sub-watersheds and the variations of ET flow may cause the variation of other components flow.

A evapotranspiration (ET) model based GIS using LANDSAT data and MODIS data with improved resolution

In this work, we integrate a popular remote sensing technique with ArcGIS to build a tool bar, named rGIS-ET, for estimating regional evapotranspiration (ET) from Landsat data and MODIS data with improved resolution. The development of rGIS-ET enables quick processing of large amount of remote sensing and other spatial data. It also provides user-friendly interfaces for modelling, output display and result analyses. Both surface temperatures and albedo were key parameters for calculating ET using Surface Energy Balance Algorithm. We adopted algorithms for estimating surface temperatures and albedo of winter wheat and summer maize field from MODIS data at 250 m spatial resolution to improve the resolutions of ET map of MODIS. We apply improved rGIS-ET to eight plain counties of Shijiazhuang city, a typical agricultural region in North China Plain, to demonstrate its utility for calculating regional ET and evaluating agriculture water resource usage. The improved ET map of MODIS could represent the spatial variation of crop ET much better than that without improved.

Knowledge-based land use mapping integrated GIS and time-series MODIS NDVI data in Lhasa River basin, Tibet

Tibet Plateau plays an important role in global changing and ecosystem studies because of its unique geographical location and topography. Lhasa river basin which locates in the center of Tibet Plateau is a typical and important region for agriculture and stockbreeding in Tibet. In this study a method of land cover mapping from 250m MODIS (Moderate Resolution Imaging Spectroradiometer) product Normalized Difference Vegetation Index (NDVI) MOD13Q1 data is presented. This knowledge-based method combines phenophase character of plants with time-series remote sensing data and Geographic Information System spatial analysis. A quality assessment analysis is performed to time-series data by temporal and spatial interpolation of invalid and missing data. The NDVI value is converted into a relative NVDI to avoid the misclassification arising by data change of spatial and temporal. The preliminary results are compared with both field observation points and classification mapped from Landsat TM imagery. The comparison indicates the result of classification is promising.

A GIS-based remote sensing model for mapping evapotranspiration distribution in a semi-arid mountain region: a case study in Taihang Mountain, North China

The evapotranspiration (ET) is one of the most important components of the water cycle in semi-arid Taihang Mountain region of North China. The spatial distribution and seasonal variation of ET will directly impact the stream flow volume and the amount of lateral recharges to the aquifers of mountain front plain. Due to significant changes in topography, the ET of this semi-arid region tends to vary dramatically both in time and space, which renders the accurate estimation of yearly or seasonal ET a difficult task. In current study, based on rGIS-ET v1.0, a regional ET model, by adding module of adjusting surface temperature in terrain, solar radiance terrain correction, and shaded relief, we improved the rGIS-ET a remote sensing model on ArcGIS platform for mapping ET distribution in such a semi-arid mountain area. With DEM of 30 meter and climate data, we run the model to estimate daily ET in mountain area using Landsat data and MODIS data, respectively. The results of model application shows that model could correct the errors of ET value caused by elevation and terrain significantly while Landsat data was used. While MODIS data was used, the model could not do terrain correction accurately for MODIS has a low spatial resolution, but MODIS data with a high temporal resolution could be used to estimate the temporal variation of ET in a mountain area.

Zonal calculation for large scale drought monitoring based on MODIS data

Temperature vegetation dryness index (TVDI) is a simple and effective methods for drought monitoring. In this study, the statistic characteristics of MODIS-EVI and MODI-NDVI at two different times were analyzed and compared. NDVI reaches saturation in well-vegetated areas while EVI has no such a shortcoming. In current study, we used MODIS-EVI as vegetation index for TVDI. The analysis of vegetation index and land surface temperature at different latitudes and different times showed that there was a zonal distribution of land surface parameters. It is therefore necessary to calculate the TVDI with a zonal distribution. Compared with TVDI calculated for the whole region, the zonal calculation of TVDI increases the accuracy of regression equations of wet and dry edge, improves the correlations of TVDI and measured soil moisture, and the effectiveness of the large scale drought monitoring using remote sensing data.