David Dent

Recent Land Degradation and Improvement in China

Abstract Land degradation is a global development and environment issue that afflicts China more than most countries in terms of the extent, economic impact, and number of people affected. Up-to-date, quantitative information is needed to support policy and action for food and water security, economic development, and environmental integrity. Data for a defined, recent period enable us to distinguish the legacy of historical land degradation from what is happening now. We define land degradation as long-term decline in ecosystem function and productivity and measure it by remote sensing of the normalized difference vegetation index (NDVI), the greenness index. NDVI may be translated to net primary productivity (NPP). Deviation from the norm serves as a proxy assessment of land degradation and improvement—if other factors that may be responsible are taken into account. These other factors include climate, which may be assessed by rain-use efficiency and energy-use efficiency. Analysis of the 23-year Global Inventory Modeling and Mapping Studies (GIMMS) NDVI data reveals that, in China over the period 1981–2003, NPP increased overall, but areas of declining climate-adjusted NPP comprise 23% of the country, mainly in south China. About 35% of Chinas population (457 million out of 1 317 million) depend on the degrading land. Degrading areas suffered a loss of NPP of 12 kgC ha−1 y−1, amounting to almost 60 million tC not fixed from the atmosphere; loss of soil organic carbon from these areas is likely to be orders of magnitude greater. There is no correlation between land degradation and dry lands; it is more of an issue in cropland and forest: 21% of degrading land is cropland and 40% is forest, 24% of the arable and 44% of the forest, respectively. There is no simple statistical relationship between land degradation and rural population density or poverty. Most identified land degradation is in the south and east, driven by unprecedented land-use change.

Quantitative mapping of global land degradation using Earth observations

Land degradation is a global issue on par with climate change and loss of biodiversity, but its extent and severity are only roughly known and there is little detail on the immediate processes – let alone the drivers. Earth-observation methods enable monitoring of land degradation in a consistent, physical way and on a global scale by making use of vegetation productivity and/or loss as proxies. Most recent studies indicate a general greening trend, but improved data sets and analysis also show a combination of greening and browning trends. Statistically based linear trends average out these effects. Improved understanding may be expected from data-driven and process-modelling approaches: new models, model integration, enhanced statistical analysis and modern sensor imagery at medium spatial resolution should substantially improve the assessment of global land degradation.

Applications of NDVI for Land Degradation Assessment

In the late 1960s, several researchers began using red and near-infrared reflected light to study vegetation (Pearson and Miller 1972). In the late 1960s, ratios of red and near-infrared light were used to assess turf grass condition and tropical rain forest leaf area index (Birth and McVey 1968; Jordan 1969). Compton Tucker was the first to use it for determining total dry matter accumulation, first from hand-held instruments (Tucker 1979), and then from NOAA AVHRR satellite data (Tucker et al. 1981, 1985), demonstrating that the growing season integral of frequent NDVI measurements represented the summation of photosynthetic potential as total dry matter accumulation. Starting in July 1981, a continuous time series of global NDVI data at a spatial resolution of 8 km has been available from the AVHRR instrument mounted on NOAA weather satellites. Soon, researchers realized the value of NDVI time-series remote sensing (Goward et al. 1985; Justice et al. 1985; Townshend et al. 1985; Tucker et al. 1985). This early work was the spur for development of the higher-resolution Moderate-Resolution Imaging Spectroradiometer (MODIS) instrument. The application of satellite NDVI data has blossomed into many fields of natural resources investigation (see Annex 1). One particular appeal of remote sensing in the study of large geographic areas, or at multiple times over the year(s), is the potential for cost savings (Pettorelli 2013). We examine the use of NDVI in research on land-use and land-cover change, drought, desertification, soil erosion, vegetation fires, biodiversity monitoring and conservation, and soil organic carbon (SOC).

Key Issues in the Use of NDVI for Land Degradation Assessment

A substantial body of research has established the correlation between NDVI and aboveground biomass, and knowledge of the theoretical basis for using satellite-derived NDVI as a general proxy for vegetation conditions has advanced (Mbow et al. 2014; Pettorelli et al. 2005; Sellers et al. 1994). Reduction of primary productivity is a reliable indicator of the decrease or destruction of the biological productivity, particularly in drylands (Wessels et al. 2004; Li et al. 2004). NPP expressed in g of C m−2 years−1 and quantifies net carbon fixed by vegetation. According to Cao et al. (2003), NPP is “the beginning of the carbon biogeochemical cycle,” defined mathematically as in Eq. (5.1):

Humus – Guardian of Fertility and Global Carbon Sink

\mathrm{N}\mathrm{P}\mathrm{P}=f\left(\mathrm{NDVI,,,,PAR,,,,fPAR,,,,aPAR,,,,LAI}\right)

Main Global NDVI Datasets, Databases, and Software

where fPAR is the fraction of absorbed photosynthetic active radiation, aPAR is the absorbed photosynthetic active radiation, and LAI is the leaf area index. Changes in NPP or, rather, its proxy NDVI induced by land degradation can be measured using a range of remote sensing techniques so remote sensing has become an essential tool for global, regional, and national studies of land degradation (Anyamba and Tucker 2012; Bai et al. 2008; Bajocco et al. 2012; de Jong et al. 2011b; Field et al. 1995; Horion et al. 2014; Le et al. 2014; Prince and Goward 1995). Many approaches have been developed to estimate NPP, notably the Global Production Efficiency Model (GLO‐PEM) (Prince and Goward 1995), the Light-Use Efficiency (LUE) Model (Monteith and Moss 1977), the Production Efficiency Approach (Goetz et al. 1999; Goward and Huemmrich 1992), and the Sim‐CYCLE (Ito and Oikawa 2002). And models have been developed to estimate NPP directly from remotely sensed NDVI at a global scale. Running et al. (2004) offered Eq. (5.2):

The Soil Cover

\mathrm{N}\mathrm{P}\mathrm{P}=\varSigma \left(\varepsilon \times \mathrm{N}\mathrm{DVI}\times \mathrm{P}\mathrm{A}\mathrm{R}-{\mathrm{R}}_{lr}\right)-{\mathrm{R}}_g-{\mathrm{R}}_m