The soil salinity map derived from this updated Soils database is presented in Fig.5and is available on the FAO website.This updated information was largely needed to plan for land use changes that came about because of rising urban cities and growing rural populations, and to curb associated land degradation by erosion, pollution, salinity, as well as biodiversity losses.More recently, FAO through the Intergovernmental Technical Panel on Soils published the Status of the World’s Soil Resources report , intended to serve as a reference document on the status of global soil resources to support studies of regional assessment of soil change.It also contains a synthesis report for policy makers that summarizes its findings, conclusions, and recommendations.The SWSR report identifies the likely rapid increase of salt-affected soils globally and estimates that currently each year some 0.3–1.5Mha of farmland is taken out of production because of soil salinity problems.The SWSR report also states that about half of the total currently salt-affected soils are further decreasing their production potential.Annual economic costs were estimated to be about US $440 per ha of salt-induced agricultural land.Currently available maps continue to be out-of-date and too coarse for predicting trends on soil salinization.Global estimates of salinization combine different regional estimates that are not necessarily compatible.It is already noted that percentages vary widely between various literature sources.Across the world,fodder growing system countries and regions typically apply different soil classification systems, and as a result the definition of saline or sodic soils varies, thus changing the acreage of salt-affected lands.

A harmonized soil salinity classification system is needed that is universally applied.Gathering accurate, up-to-date information is critical for developing policies to halt the trend of increasing soil salinity across the world and regionally.Efforts to develop an updated and harmonized global soil salinity map were recently initiated by FAO through the Global Soil Partnership or GSP , through mapping of soil EC, SAR, and pH using existing country-level data.Soil salinity and the increase in areal extent is a serious global threat to agricultural production as soil degradation jeopardizes reaching a food-secure world.The only database that currently provides soil salinity data with global coverage is the Harmonized World Soil Database, but it is outdated and has several limitations when assessing changes in soil salinity and its areal extent.Except for a few country-focused reports, there is limited information on the world’s changing extent of salinized soils.Therefore, we recommend taking steps toward a new assessment.There are various reasons to suggest that the areal extent of soil salinization is increasing as well as becoming more severe.Information on such trends is extremely relevant as global and national policies on land use are being developed to advance Sustainable Development Goalsand to mitigate and/or adapt to climate change.Moreover, areas of salt-affected irrigated lands are inconclusive and vary between 25% and 50%depending on the data source.Soil salinization may be accelerating for several reasons including the changing climate.Rising temperatures increase soil evaporation and crop water requirements, enhancing soil salinization in areas already prone for salinity.Especially, coastal regions will be subjected to increasing risk of salinization by rising seawater levels, thereby pushing more saltwater into coastal aquifers, and increasing groundwater salinity.In addition, the likelihood of extreme storms and tsunamis can cause flooding of seawater, resulting in saltwater infiltration into soils and contaminating groundwater resources.

In his analysis of climate change impacts on soil salinization processes, Corwin states that the consequences of climate change have been overlooked and that changes in soil salinity extent will need to be monitored and mapped.He suggests that both proximal and remote sensors are the best methods to achieve this in a timely manner.Another reason that the area of saline soils is expanding relates to the increased use of marginal waters for irrigation, as decreased freshwater availability encourages application of treated wastewater or low salinity water for irrigation.Also, changing land uses from prime agricultural land to residential development promotes cultivation of more marginal lands, thereby enhancing the potential for land degradation.Furthermore, the decreasing availability of freshwater promotes more efficient irrigation methods such as drip and sprinkler irrigation, leading to reduced leaching of accumulated soil salts in regions with limited winter rains.Yet, to meet the world’s demand for nutritious food with the rising population, one may expect a further increase in irrigated area, especially in regions where freshwater availability is adequate.Lastly, salts accumulate over extended periods of continuous irrigation, thus further causing more salinity-prone areas over time.A universal global soil salinity map can be achieved using satellite imagery, soil properties maps, other land surface information, and advanced data analysis methods such as machine learning techniques.A recent example of such an approach was taken by Ivushkin et al., supported by the International Soil Reference and Information Centre.In their work, a total of six soil salinity maps were produced for 1986, 2000, 2002, 2005, 2009, 2016, using thermal IR imagery data from Landsat satellites.Their analysis presented a clear trend over this 20-year period, indicating that the global area of salt-affected soils increased from about 900 to 1000Mha, at an annual rate of about 2–5Mha/year.Various limitations of their methodology were given, including the need for higher spatial resolution, more ground truth data for regions with sparse data, uncertainty associated with temperature response due to plant variations in salt tolerance, and potential improvement using machine learning techniques.Salt-affected soils have significant impacts on the environment, freshwater availability, and agricultural production.Updated maps are needed to quantify soil salinization rates and to inform country level and new international policies and strategies to protect soils from further salinization.We urge prioritizing development of remote sensing instruments for future satellite missions that focus on observing spatial and temporal changes in land degradation, including soil erosion and salinity, at a global scale.

Detecting and monitoring soil salinity across agricultural regions is needed for inventorying soil resources; for identifying trends and drivers in salinization; and for judging the effectiveness of reclamation and conservation programs.Due to the impracticality of directly measuring root zone ECex over large areas , most regional-scale salinity assessment research has focused on alternative measures of salinity obtained through aerial photography and satellite remote sensing.Despite being developed many decades ago, remote detection of salinity has not been widely used in salinity monitoring programs and has achieved only limited success to date.However, methodological and technological advances made over the last 20 years suggest the routine use of remote sensing for monitoring agricultural salinity may be possible.Two approaches to remote salinity detection have been used: indirect and direct.With indirect methods, the level of root zone salinity is inferred based on crop growth and health,chicken fodder system usually as indicated by canopy spectral reflectance or thermographic data.The reflectance of certain visible or infrared spectra generally differs for healthy and stressed leaves.Thus, if a correlation between root zone ECex and spectral response can be established, regression or classifier models can be developed to quantify or label soil salinity levels in a remote sensing image.Direct methods detect salinity in bare soils based on the reflectance properties of surface salts and crusts.Sections of landscapes with and without surface salts can be distinguished due to the high reflectance of salt covered areas in the visible part of the spectrum.Within salt covered areas, salinity levels and salt types may be differentiated because of the effects that salt abundance, mineralogy, moisture, color, and surface crusting and roughness all have on reflectance.The direct approach is useful for assessing salt marshes and other highly saline, non-agricultural landscapes, as well as for tracking encroachment or appearance of barren, high salinity areas in dryland pastures and range lands.However, it has less utility for agricultural regions because of the presence of extensive vegetation.Therefore, we focus on indirect RS methods for soil salinity monitoring.By the middle of the 20th century, aerial photography and image analysis were touted as a means of inventorying crops and detecting disease.Portable or airborne spectral reflectance instruments did not exist, but laboratory measurements made on tissues from leaves in varying states of distress could reveal, for a given crop and development stage, the portion of the spectrum most sensitive to variations in leaf health.Aerial photographs sensitive to the identified spectral range could then be made using an appropriate combination of film and lens filter.

Through analysis of the aerial images, it was proposed that areas with healthy and diseased plants could be distinguished.Myers et al.were the first to connect aerial images of crops with root zone salinity.Working in Texas cotton fields, Myers et al.found that the salinity level in the 0.3–1.2m soil layer could be correlated with the spectral reflectance of cotton leaves, determined from aerial photographs using infrared film and a dark red filter that was sensitive at 675–900μm wavelengths.In a subsequent paper, Myers et al.reported it was possible to distinguish five levels of salinity and to estimate with reasonable accuracy the degree of salinity in the soil profile.It was also found that soil salinity could be predicted with reasonable accuracy from leaf temperatures measured with an infrared radiometer.Thomas et al.examined in greater detail the spectral reflectance of salt-affected cotton leaves and found that they changed during the growing season.At most wavelengths, percent reflectance from individual leaves was negatively correlated with salinity early in the year and positively correlated later.Multiple regression analyses of aerial image density indicated that under field conditions reflectance was influenced by soil salinity and percentage ground cover.The Landsat program and launch of the first operational Landsat satellite in 1972 spurred interest in using multi-spectral satellite imagery for natural resource management.Notable early examples of using space borne aircraft to detect salinity include identifying salt flats in Imperial Valley, California from photo images taken aboard Apollo 9and distinguishing saline from non-saline rangelands in South Texas using Skylab satellite imagery.The review of Metternicht and Zinck covers advances made during this period with respect to direct observation of visible surface salts.With the growing availability of multi-spectral reflectance data from satellites and other platforms, it became common from the 1970s onward to quantify multi-band canopy reflectance using vegetation indices such as the Normalized Difference Vegetation Index, NDVI¼/ , where R and NIR are spectral reflectance in the visible red and near-infrared bands, respectively.Wiegand et al.used imaging data from the SPOT-I satellite to evaluate the relationship of NDVI and the Greenness Vegetation Index to plant growth and yield in a single salt-affected, irrigated cotton field in Texas.Later, Wiegand et al.determined NDVI and GVI for four cotton fields in San Joaquin Valley , California using airborne photographic imagery made with multiple lens filters.Regression equations with NDVI and GVI as predictor variables were used to estimate salinity at about 100,000 pixels per field.The last 2 decades have seen a steady increase in the availability of remote sensing data, in the capabilities of various sensors and platforms, and in remote sensing applications.Even with improved technologies, a major problem with indirect salinity detection methods is that a single image generally cannot differentiate salinity-induced crop stress from stress caused by other factors such as weather, pests, and water management.Lobell et al.addressed this difficulty by evaluating multi-year data, hypothesizing that soil salinity is relatively constant compared to other more transient stressors.Lobell et al.found that using 6 years of reflectance data greatly improved the correlation between salinity and wheat yield, whereas Lobell et al.successfully evaluated regional-scale salinity using a 7-year average enhanced vegetation index derived from satellite MODIS data.Multi-temporal data was also used by Caccettaand Furby et al.for improved soil salinity classifications.Along the same lines, Zhang et al.used interpolated and integrated vegetation index time-series data as an explanatory variable rather than analyzing single-date data.Whitney et al.later applied the same integrated index method to the SJV and concluded that multi-year data further enhanced correlations with soil salinity.The use of environmental covariates as additional predictor variables in regression equations and classifiers has also improved accuracy.Scudiero et al.developed a linear regression equation for estimating soil salinity using spatial precipitation and temperature data, croptype data, and multi-temporal Landsat 7 ETM+ canopy reflectance data.They calibrated their model using data for thousands of Landsat 7 pixels at 30m resolution across 22 fields for which ground truth salinity data were available.For each 30 30m Landsat pixel, average root zone ECe for a 6-year period was modeled using the Canopy Response Salinity Index, CRSI, which combines spectral reflectance in the green, blue, red, and near-infrared bands.