Promoting the use of isotopic techniques to combat soil erosion: An overview of the key role played by the SWMCN Subprogramme of the Joint FAO/IAEA Division over the last 20 years

The International Atomic Energy Agency (IAEA), through the Joint Division with the Food and Agriculture Organization (FAO) of the United Nations, assists its Member States in applying nuclear techniques to alleviate challenges in food safety, food security and sustainable agricultural development. The Soil and Water Management & Crop Nutrition (SWMCN) Subprogramme, within the Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture, has made significant contributions to the development of isotopic techniques for the assessment of soil degradation and the development of efficient soil and land conservation approaches. These techniques include fallout radionuclides such as 137Cs, 210Pbex, 7Be, and 239+240Pu as well as 13C stable isotope and compound‐specific stable isotope analyses. These methodologies were developed and/or refined through the work of researchers from developed and developing countries who were selected to work within the frame of IAEA's Coordinated Research Projects (CRPs). Internal research activities implemented in the Joint FAO/IAEA's SWMCN Laboratory in Seibersdorf supported the work accomplished in the CRPs.


| INTRODUCTION
A major challenge for the global population is to ensure food security. It is projected that the world population will exceed 9 billion by the year 2050, and therefore food production would have to be doubled to fulfil this requirement. This need will occur principally in developing countries, where most people rely on agriculture for their livelihood (Food and Agriculture Organization [FAO], 2015[FAO], , 2017United Nations Environment Programme, 2015).
Model projections suggest that global surface temperature at the end of the 21st century is likely to exceed 1.5°C relative to the period 1850-1900 (Intergovernmental Panel on Climate Change [IPCC], 2013). The latest 2014 report produced by the IPCC stresses the projected impact of climate change, highlighting hazards, vulnerability, and exposure resulting from these changes (IPCC, 2014). To address these critical issues of climate change and population growth, the FAO and the global scientific community are helping farmers to develop climate-smart agricultural practices and systems, which can adapt to the impacts of climate change and variability (i.e., climate change adaptation). Although having the potential to increase food production, these practices also improve soil and water management and reduce land degradation, and associated soil carbon loss and greenhouse gas emissions (i.e., climate change mitigation; FAO, 2013).
Land degradation and soil erosion by water are associated with the irreversible loss of fertile soil, reduced soil productivity, increased siltation and pollution of water bodies.
The annual global cost of soil erosion and its associated downstream sedimentation were established to be approximately U.S. $400 billion (Pimentel, 2006). Land degradation also increases CO 2induced climate change through release of CO 2 (WMO, 2005). In particular, by significantly reducing soil quality and fertility, soil erosion may contribute to intensifying the impact of climate change. Erosion leads to soil disruption and aggregate breakdown and, hence, partial mineralization of soil organic carbon and CO 2 release (Lal, 2004;Lal, Griffin, Apt, Lave, & Morgan, 2004;Lal & Pimentel, 2008;Ran, Lu, & Xin, 2014;Reichstein et al., 2013).
Current social and economic impacts related to soil degradation have generated a pressing need for obtaining accurate information on soil erosion magnitudes affecting agricultural land to strengthen soil conservation strategies.
The main assumption for the use of these radiotracers in soil erosion and sedimentation investigations is similar. When deposited on the soil surface, they are strongly bound to soil colloid particles (clay and organic matter) and move in the environment by mechanical processes such as soil erosion (water, wind, or tillage erosion). To use FRN methods, it is necessary to establish the content of FRNs at a reference location (i.e., an undisturbed site where neither significant erosion nor deposition is occurring) located near the study area.
Investigated areas are identified as erosional or depositional by comparing their FRN inventories with those of the reference site. Levels below the reference site value will indicate soil eroded, and those in excess indicate soil deposited (see IAEA, 2014;Mabit, Benmansour, & Walling, 2008;Zapata & Nguyen, 2009 To ensure sustainable agricultural management, there is a need not only to quantify soil erosion rates but also to localize the sources of land degradation in the landscape, to identify appropriate conservation measures, and to test and assess their efficiency. ranging from a few rain events ( 7 Be) to circa 60 years ( 137 Cs and 239+240 Pu) and up to some 100 years for 210 Pb ex (Mabit, Benmansour, & Walling, 2008). When compared with those of conventional measurement, soil redistribution rates derived from FRNs require only one field campaign, making this approach time and resource efficient.
Like all other techniques or approaches, FRN-based methods have their strengths and limitations. Besides, they are based on some key assumptions (i.e., uniform initial spatial distribution, and fast and almost irreversible adsorption to fine soil particles) that need to be met to ensure the reliability of the results generated. Using FRNs requires that at each sampling point, the total isotope inventory is measured, to ensure appropriate comparison of the study site with the reference site. The selection of appropriate reference sites unaffected by erosion processes is also crucial. An in-depth understanding of particle sorting during erosion and deposition phases is important when interpreting FRN redistribution data, as FRNs are fixed preferentially to fine particles. Otherwise, soil loss rates could be overestimated and deposition rates underestimated. The analysis of FRNs in soils and sediments requires specialized equipment (e.g., gamma spectrometry for 137 Cs and 7 Be, gamma spectrometry and/or and alpha spectrometry for the determination of 210 Pb ex ) as well as trained staff.
By essence, all assessment methods-even the more mature ones -have shortcomings that can be improved upon. For example, the pros and cons of the 137 Cs method have been discussed and debated within the scientific community Parsons & Foster, 2011), and, as constructive follow-up, pragmatic proposals to refine this method have been suggested (e.g., Zhang, Zhang, & Wei, 2015).
As mentioned by Mabit et al. (2013) and confirmed by Zhang et al. (2015), a sound sampling design is crucial when using this method to obtain reliable soil erosion rate estimates. In addition to scaling issues, comparison with direct measurements and monitoring methods should always be performed considering the processes involved by the different methods and the time scales investigated. Moreover, the major advantages and disadvantages of FRNs were already discussed in detail both in several review papers focusing on 137 Cs (Mabit, Benmansour, & Walling, 2008;Walling & Quine, 1991), 7 Be (Taylor et al., 2013), 210 Pb ex , and 239+240 Pu (Alewell et al., 2017) and in a comprehensive handbook summarizing the experience of last decades of scientific activities conducted by using these radioactive tracers (IAEA, 2014).
For this reason, the detailed methodological assumptions and limitations of FRN methods will not be presented here, as the objective of this paper is to provide an overview of the involvement of IAEA in the development of FRN methods over the last decades. to quantify concretely soil conservation effectiveness for ensuring sustainable local and regional land management (see Figure 2).

| CONFIRMING AND IMPROVING THE FRN METHOD
The use of nuclear techniques (especially the 137 Cs method) for erosion and sedimentation assessments started in the 1960s and underwent, in the 1980s and 1990s, rapid methodological development that led to comprehensive agro-environmental evaluation tools.
A major milestone was achieved at Exeter University (UK), with the development of conversion models for translating variations of 137 Cs inventories into soil erosion/deposition rates (Walling & He, 1999;Walling & Quine, 1990).
In the mid-1990s, the IAEA became actively involved in the development of nuclear techniques through soil loss and sedimentation investigations (IAEA, 1995(IAEA, , 1998. The FRN method thus began to be used routinely in a large number of situations worldwide and was focused on a wide range of research topics such as estimation of rates of particular erosion processes (such as gully erosion, tillage erosion, and wind erosion), assessment of land use impact, testing of conservation measures, and validation of erosion models and sediment chronology in reservoirs and on alluvial plains.
However, the major achievement of the two CRPs was an overall standardization of the methodological approaches used by different research teams. Standardized methods and protocols were published in a widely accessible handbook edited by Zapata (2002). This handbook focuses on the 137 Cs method but also provides information on 210 Pb ex and 7 Be. It also highlights methodological approaches that are needed for both erosion and sedimentation studies such as: • guidelines on site selection and sampling design (Pennock & Appleby, 2002a); • sampling and sample pretreatment (Loughran, Wallbrink, Walling, & Appleby, 2002;Pennock & Appleby, 2002b); • gamma spectrometry for FRN measurements ; • assessment of spatial distribution of 137 Cs (Loughran, Pennock, & Walling, 2002); • use of FRN conversion models for soil redistribution studies (Walling, He, & Appleby, 2002); • specific use of 137 Cs in Chernobyl-affected areas (Golosov, 2002), and • 137 Cs in situ measurements and 7 Be and 210 Pb ex methodologies . of conversion models for all three radionuclides Walling, Zhang, & He, 2011). This new set involved six improved models for 137 Cs conversion as well as models adapted for the use of 210 Pb ex and 7 Be:
The D1.50.08 CRP brought also several findings on technical challenges, constraints, and limitations of FRN methods. One of the major challenges is the selection of a suitable reference site. In some areas, the problem can be caused by the high variability of rainfall (Belyaev et al., 2009), but in most cases, the problem originates from the lack of undisturbed land in steep mountains, densely populated regions, and intensively urbanized areas (Haciyakupoglu et al., 2005;Mabit et al., 2009). Other important limitations are low inventories of 137 Cs in the Southern Hemisphere and high heterogeneity of 210 Pb ex inventories (Belyaev et al., 2009;Mabit et al., 2009). Among the technical (analytical) challenges, variability in the capacity to measure 210 Pb ex precisely was raised, as a proficiency test highlighted that only 31% of laboratories participating in the CRP achieved sufficient accuracy for 210 Pb ex activity measurement (Shakhashiro & Mabit, 2009).
The major methodological principles and features of the FRN methods, as well as their advantages and limitations, were discussed in several papers Mabit, Benmansour, & Walling, 2008;Taylor et al., 2013;Zapata & Nguyen, 2009). However, some of these limitations and technical difficulties can be overcome by either more extensive sampling, to cover the variability Mabit, Bernard, et al., 2008), or longer counting times to reduce error for low activity samples and more careful data interpretation.

| INTEGRATING SOIL AND WATER MANAGEMENT AT THE WATERSHED SCALE
For promoting their sustainable use, natural resources should be managed in an integrated way at the watershed scale (IAEA, 2014).
Although some early work was undertaken in areas ranging from some hectares to a few square kilometres under D1.50.08 CRP and elsewhere, the need to develop more comprehensive watershed studies combining an oriented sampling strategy and a geostatistical approach, geographic information system, and 137 Cs data set information (e.g., Navas, Machín, & Soto, 2005) had already emerged.
Evaluating the magnitude of soil erosion over large areas is highly complex. Therefore, to gain better knowledge of sediment dynamic at Under this CRP, a harmonized protocol was also developed for the application of CSSI techniques using FAs as specific organic compounds, to localize sediment source and erosion hot spots at the watershed scale in a range of agro-environments (Gibbs, 2014). By linking fingerprints of specific land use to suspended or deposited sediment, CSSI techniques have proven to be a highly innovative approach for establishing the source of eroded soil or transported sediment. In addition, it is particularly useful for identifying areas delivering high sediment loads and thus contributing to water pollution (Gibbs, 2014;Heng, Sakadevan, Dercon, & Nguyen, 2014). CSSI-based techniques provide information on sources and can provide quantitative information as well as an additional dimension to generic source information that is more relevant to land use management decision making. However, this technique is still in its infancy with only a few published applied studies using specifically δ 13 C of FAs to explore sediment transfer and origin within agro-ecosystems. Several aspects are still under investigation such as the best proxy or parameter to be used to convert isotopic signature into soil proportion.
As FRNs have proven to be powerful tools for assessing landscapewide soil redistribution and identifying erosion processes, their integration with CSSI analysis opened new opportunities for improving areawide soil conservation strategies for agricultural landscapes (Dercon et al., 2012;Heng et al., 2014 showed how past land degradation and its link with land use history over the last centuries can be revealed (Gibbs, 2014). The combined uses of FRNs and CSSI have been tested successfully in Australia (Hancock & Revill, 2013) to identify hot spots and reassess previous understandings of landscape sediment dynamics (e.g., relative importance of channel bank erosion and other sources). Similarly, the combined use of CSSI and geochemistry has been used successfully to identify the importance of damaged pastures as erosion hot spots in the U.K. (Blake, Ficken, Taylor, Russell, & Walling, 2012).
In addition, an innovative and accurate sampling system, the Fine Increment Soil Collector, was developed at the SWMCN Laboratory to facilitate the precise determination of soil depth distribution of radionuclides . Ryken et al. (2016) reported that the Fine Increment Soil Collector, compared with other existing sampling devices, allows for more precise collection of soil for erosion FRN-derived studies.

| USE OF ISOTOPIC AND NUCLEAR TECHNIQUES TO ADDRESS SOIL EROSION IN THE CONTEXT OF CLIMATE CHANGE
In several regions across the globe, climate change is impacting the precipitation regime (Christensen et al., 2007), resulting in increased drought periods and high-intensity rainfall events (IPCC, 2013(IPCC, , 2014. As a result of climate variability and global warming, world average soil loss is predicted to further increase significantly (Li & Fang, 2016;Yang, Kanae, Oki, Koike, & Musiake, 2003). Soil erosion decreases soil productivity through soil, nutrient, and organic matter losses; deterioration of overall soil health; decrease of fertility, production potential, and biological activity; breakdown of soil structure; increase of soil erodibility; and reduction of soil water holding capacity. Poor soil quality further accelerates soil erosion, in particular on steep farmland, where this process is intensified by overgrazing and improper agricultural practices (McHugh, 2007;Tiwari, Sitaula, Bajracharya, & Borresen, 2009;Valentin et al., 2008). The intensification of upland erosion also increases sediment delivery downstream, causing further problems such as off-site erosion effects among which the most important is the siltation in water reservoirs and pollution of water sources and coastal sea waters resulting in the dying of coral reefs (e.g., Smith & Wilcock, 2015).
In line with these statements, an expert meeting on "Soil and water conservation for climate change adaptation in agricultural uplands" was held in December 2014. This meeting paved the way for the new CRP D1.50.17 on "Nuclear techniques for a better understanding of the impact of climate change on soil erosion in upland agro-ecosystems" (2016-2021) that focuses on the refinement and development of isotopic techniques for improving our knowledge of the impact of climate variability on upland agricultural areas, so that better management can ensure sustainable production systems that will be resilient to the impacts of climate change.
A range of nuclear and isotopic techniques are being used to support these goals, including FRNs ( 137 Cs, 210 Pb ex , 7 Be, 239+240 Pu) and CSSI techniques as well as the cosmic-ray neutron probe. Combinations of these mature or innovative techniques are being tested as indicators of changes in soil and water resources occurring in fragile upland areas and for unravelling the relative importance of climate variability and agricultural practices.
Within this project, a newly proposed artificial FRN, that is, 239+240 Pu, is being further tested and validated as a soil redistribution tracer. The Pu isotopes have several advantages over the wellestablished 137 Cs method, namely, a more homogenous spatial distribution that is unaffected by nuclear power plant accidents and a long half-life, which ensures a much longer environmental availability than that of 137 Cs (Alewell et al., , 2017Hoo, Fifield, Tims, Fujioka, & Mueller, 2011;Lal, Tims, Fifield, Wasson, & Howe, 2013;Tims, Everett, Fifield, Hancock, & Bartley, 2010).
To differentiate the influence of climate variability and agricultural practices, paired catchment and FRN resampling approaches will be tested within the CRP activities. The paired catchment concept consists of investigating two small catchments of comparable size having similar geomorphological and climatic condition. One of these catchments should be native or protected with limited to no anthropogenic activities (e.g., a natural park), and the other one should host agricultural activities. Both areas should be investigated using FRNs and CSSI techniques such that results can be compared. A slightly modified approach can be proposed in focusing on the FRN dating aspect complemented by CSSI techniques on the sediment of both watersheds deposited in reservoirs.
A FRN resampling option, using 137 Cs (Loughran & Balog, 2006;Porto et al., 2014) or even 210 Pb ex as suggested recently by Porto, Walling, Cogliandro, and Callegari (2016), could be an easier and more valuable strategy for distinguishing and effectively apportioning the impact of climate change in upland agro-ecosystems. However, to test this approach in the context of the CRP and its specific objectives, it is mandatory (a) that the research is performed in an area already investigated with FRNs in the past, (b) that the lapse of time between the two FRN sampling campaigns is sufficient to notice significant changes of soil redistribution magnitudes (at least one decade after the first investigation is needed), and most importantly, (c) that the land use has remained the same since the first FRN investigation.
If all these assumptions-especially the latter-are respected, it can then be expected that only the climatic variability and its impact could explain a variation (if any) in the soil redistribution magnitude.
Since the start of the CRP D1.50.17, some preliminary and promising findings contributing to refining the FRN and CSSI techniques have already been obtained and published by participants of the CRP: • the Department of Environmental Sciences (University of Basel, Switzerland) in collaboration with the SWMCN Laboratory developed a "universal" conversion model, called MODERN (i.e., Modelling Deposition and Erosion rates with RadioNuclides), to assess soil redistribution magnitudes from FRN measurements . MODERN is the only conversion model that can be used for 137 Cs, 210 Pb ex , and 7 Be as well as for the new soil tracer 239+240 Pu; • plutonium isotopes (i.e., 239+240 Pu) have been tested and validated relative to other more mature radioisotopic approaches for deriving soil erosion rates under various upland agro-environments in Switzerland Meusburger et al., 2018) and also in South Korea ; • the development of a cost-effective sampling strategy when using CSSI techniques to reduce and optimize analytical labour , and the use of artificial mixtures to confirm the accuracy and reliability of the CSSI mixing models, a new proposal for converting isotopic proportion into soil proportion using the FA concentrations instead of the total % C org has been made (Alewell, Birkholz, Meusburger, Schindler Wildhaber, & Mabit, 2016); • through their studies in South West England, Taylor, Keith-Roach, Iurian, Mabit, and Blake (2016) reported that the use of cosmogenic 7 Be as a soil erosion and/or sediment tracer requires an accurate knowledge of its temporal fallout dynamics as its deposition flux can be highly variable across months and seasons; • a study performed in the Madagascar highlands (see next section) highlighted that the combined use of 137 Cs and 210 Pb ex allowed evaluating the effectiveness of ancient terracing practices to protect soil against erosion. For the first time, this pilot Malagasy FRN investigation highlighted that despite low expected 137 Cs activity, this method can still be used with success in African countries located in the Southern Hemisphere. The suitability of 210 Pb ex as a soil tracer under Malagasy agro-climatic condition was also verified and provided similar results to those obtained with 137 Cs (Rabesiranana, Rasolonirina, Solonjara, & Mabit, 2016). 8.1 | Assessing the effectiveness of terraced agriculture in Madagascar's agricultural highlands through the use of FRNs Soil degradation-mostly due to soil erosion-is of national concern in Madagascar. According to the FAO, around one third of the island's total soil area is degraded (Nachtergaele et al., 2011) Soil erosion rates of an unprotected agricultural field and a terraced field were quantified in an experimental study area located in the eastern central highlands, 40 km east of Antananarivo (Photo 1).
As reported by Rabesiranana et al. (2016), this pioneer use of FRNs (i.e., 137 Cs and 210 Pb ex ) demonstrated that, in promoting the use of traditional terrace systems, soil erosion can be reduced by up to 40%. This Malagasy study also highlighted that traditional terracing could significantly limit the transfer of sediment and therefore the downstream potential off-site impact of the agro-ecosystems in allowing a better soil redistribution within the agricultural fields.

| FRN techniques contribute to the improvement of soil conservation in Moroccan agro-ecosystems
In Morocco, reducing on-site and off-site impacts associated with soil erosion and land degradation is a major concern for improving soil quality and protecting downstream water quality and quantity. Soil erosion is the main land degradation process in Morocco and affects at least 13% of its land area (Nachtergaele et al., 2011).
In partnership with the SWMCN Subprogramme, the Centre Case studies using 137 Cs and 7 Be were carried out in three Moroccan agricultural sites-Marchouch, Harchane, and Oued Mellah-which are located respectively in the regions of Rabat, Tétouan, and Chaouia-Ouardigha. Long-term soil erosion rates of the three regions as evaluated by the 137 Cs method ranged from 8 to 58 t·ha −1 ·year −1 (Benmansour et al., 2013). For the experimental sites in Rabat and Tétouan, the results obtained using 7 Be indicated that soil losses have been reduced appreciably under no-till as compared with conventional tillage. Thus, by adopting no-tillage soil conservation practices, soil loss from Moroccan watersheds can be diminished by approximately 40%, leading to a significant reduction of sedimentation into reservoirs. In the Oued Mellah watershed, high-density Atriplex plantations reduced soil loss by approximately 60% to 80%, whereas soil erosion decreased by 58% on sites with fruit plantations plus cereals.

| FRNs as decision-making tools to minimize land degradation in Vietnam
Lowland agricultural areas are scarce in Vietnam due to increasing population growth. Thus, there is an increasing cultivation pressure on the Vietnamese sloping lands, which is causing severe soil degradation, leading to a significant reduction of soil fertility. Approximately 40% of Vietnam's total land area is affected by soil erosion (Sanh, 2006). To compensate for nutrient losses associated with erosion processes, high levels of chemical fertilizers (>250 kg·ha −1 ) are being applied to maintain current crop yields (Hai & Dung, 2014 Prior to adopting soil conservation measures, the annual soil erosion rates ranged from 12 to 42 t·ha −1 ·year −1 for different land slopes, with a magnitude above 40 t·ha −1 ·year −1 observed on 25°to 35°s lopes. These high soil erosion rates led to significant losses of organic matter and essential plant nutrients. For example, annual organic matter losses of 555, 840, and 1,200 kg·ha −1 were recorded for slopes of 5°to 15°, 15°to 25°, and 25°to 35°, respectively. However, by using effective soil conservation measures, eroded areas were significantly reduced. For example, in areas with a slope of 25°, by intercropping pineapple with cashews, the yearly soil loss was reduced from 42 to 27 t·ha −1 , relative to pineapple monoculture. On slopes of 18°to 20°, contour farming, compared with the control for tea plantations, decreased soil erosion rates by 36%. Similarly, for tea plantations on slopes of 8°to 10°where 1.4-m contour lines were utilized, soil erosion was reduced by 40% (i.e., 24 compared with an original 34 t·ha −1 ·year −1 for the untreated slopes). For coffee plantations, by creating miniature catchment basins at the base of each coffee plant, soil erosion decreased by 42% compared with that of the untreated control (16 compared with 28 t·ha −1 ·year −1 ). For mulberry fields, contour farming together with intercropped maize was even more effective, with soil loss decreasing by 54% (11 compared with 23 t·ha −1 ·year −1 ). For vegetable fields, by the use of terraced farming on slopes of 7°to 10°, the soil erosion rate decreased by 59% (7 compared with 17 t·ha −1 ·year −1 ). Preliminary tests using CSSI techniques have Photo 2. Field visits of key stakeholders and decision makers to a coffee plantation which had implemented soil conservation strategies (Lamdong Province, Vietnam) (© Mohammad Zaman, IAEA) highlighted the fact that coffee plantations without soil conservation measures were the major source of eroded soil in the study area.
Assuming that the 13 million ha of sloping lands in Vietnam contain 1% soil organic matter (presuming a 5% nitrogen and 0.5% phosphorus content) and that the average soil erosion rate is 25 t·ha −1 ·year −1 , then a 47% reduction in soil erosion will allow the retention of soil nitrogen and phosphorus with estimated fertilizer cost values of U.S. $55 million for nitrogen and US $19 million for phosphorus.

| CONCLUSION
During the last 20 years, the Joint FAO/IAEA Division, through various activities conducted by the SWMCN Subprogramme, has significantly contributed to the development and refinement of efficient isotopic techniques to localize degraded agricultural areas, to evaluate soil redistribution magnitudes from field to watershed scales, to assess the effectiveness of soil conservation strategies, and, more recently, to establish the origin of the mobilized and deposited sediment.
Complementing conventional soil erosion measurement methods and modelling, FRN and CSSI techniques have deepened our understanding of soil erosion and its related agro-environmental impacts.
The SWMCN Subprogramme has disseminated, implemented, and/or supported their use in more than 70 countries, particularly through the IAEA's Technical Cooperation Program resulting in the production of several national impact stories such as the ones obtained in Morocco, Madagascar, and Vietnam.
The Joint FAO/IAEA Division has, to some extent, succeeded in promoting the transition of nuclear and isotopic techniques from a validated research tool to a decision support tool. Scope for improvement remains in reinforcing the accuracy of these techniques and refining some operational parameters, but scientific evidence has demonstrated that the combined use of stable and radioisotopic tracers provides crucial information to guide decisions on the management of critically degraded areas and may support land managers to implement appropriate soil conservation measures at these hot spots. This will ultimately result in the improved efficiency of soil and water resource management, higher economic returns, and agro-environmental sustainability in the face of climate change.
This way, the work, both past and present, undertaken by the SWMCN Subprogramme has strongly contributed to the achievement of several of FAO's strategic objectives as well as the UN's Millennium Development Goals.