Comparison of conventional and artificial fallout radionuclide (FRNs) methods in assessing soil erosion

ABSTRACT

Soil erosion is a major global environmental problem. The objective of this paper is to review conventional and non-conventional-artificial radionuclide techniques for soil erosion assessment using secondary data. Data gathered reveal that accurate assessment of soil erosion rates is a pre-requirement for environmental planning and soil conservation strategies. Nevertheless, estimating rates of soil erosion is still challenging despite availability of numerous assessing methods. Many assessing methods were developed at localized scales therefore limited applications in other areas. This review compared effectiveness of different soil erosion assessing methods in different areas. Generally, soil erosion can be assessed by conventional methods and non-conventional methods e.g use of artificial fallout Radionuclide (FRNs). FRNs quantifie relatively long-term (>30 years) soil erosion and deposition, however, not suitable for short-term and individual soil erosion events. Conventional methods are associated with point data, do not provide information on spatial distribution, labor-intensive and require long monitoring periods. Assessing soil erosion method should be site-specific rather than generalizing. Therefore, there is no one take for all methods in assessing soil erosion but choice of a method to use should depend on prevailing climatic conditions, resources available and time period (short or long term) of erosion data required.

KEYWORDS:

• accuracy
• conservation
• deposition
• erosion
• fallout radionuclides
• runoff

PUBLIC INTEREST STATEMENT

Soil erosion is a major form of land degradation. The rate of erosion occurring in an area is dependent on the specific soil type and vegetation cover present. Accurate assessment of soil loss in an area is important for planning and soil erosion control strategies. However, it is still difficult to accurately assess the rate of soil loss despite that numerous assessment methods were devised. General assessment methods are usually applied when estimating soil loss resulting in inaccurate data. The inaccurate data will lead to the application of unsuitable control measures hence the persistence of the problem. Many assessing methods were developed at localized scales therefore limited applications in other areas. Soil erosion is assessed by conventional and non-conventional methods, for example use of artificial fallout radionuclide (FRN). FRNs quantify relatively long-term (>30 years) soil erosion and deposition, however, not suitable for short-term and individual soil erosion events.

Introduction

Soil erosion is the process of detachment, transportation and deposition of soil particles from land surface. It is the second most prevalent global environmental problem after population growth (ISCO, 2002). The soil erosion is referred to as “soil cancer”, because of its complexity and has caused multiple effects on human activities (Ownegh, 2003). It reduces the land value by removing the topsoil, where organic matter (OM) and nutrients are concentrated leading to reduced soil productivity (Chalise et al., 2019). Rates of loss are higher than rates of soil formation. Generally, a very long period of time (approximately 400 years) is required to replace at least 1 mm of lost soil making it a semi-nonrenewable resource (Pimentel & Burgess, 2013). Currently, the world is characterized by many soils described as fair, poor or very poor condition, and soil erosion is the major soil threats (FAO, 2015). Soil erosion is influenced by many factors that ranges from erosive to erodibility factors. It is agreed that many issues around the causes of soil erosion are still not well understood, for example whether and indeed how different countries influence the erosion of their soils (Gomiero, 2016; Rodrigo Comino et al., 2016). Current research on estimating the rates soil erosion has been focusing on identifying reciprocal relationships, for example by noting that the erosion is more severe in poor than rich countries but with little attempts to identify the causal effects. Hence, many geomorphologists see the need for new approaches and paradigms in estimating soil erosion (Gholami et al., 2018).

Previously, efforts in soil erosion were aimed at assessing rates (in units of mass/area and time) in various climatic conditions and land uses, and relied on many measurement methods (FAO, 2019; Rodrigo Comino et al., 2017). Soil erosion assessment precedes land use planning, in fact the assessment of actual and potential erosion should be part of a land evaluation in any area. Alewell et al. (2019) argued that the rates of erosion experienced in area should be measured in order to assess its impacts on the environment and conservation practices, and the implementation of conservation policies. In this situation, the land management plans have to consider both on-site and off-site impacts of erosion. The on-site impacts refer to soil loss and the decline of soil OM (SOM) content and soil structure, leading to low soil fertility and water-holding capacity, and ultimately reduced food security (Baumhardt et al., 2015). Off-site effects involve an increased flood risk and reduced lifetime of reservoirs (FAO, 2019).

Assessment of soil erosion rates requires time, effort and funds to measure the soil loss at various scales ranging from very small plots (<1 m²) to large basins (>1000 km²) (García-Ruiz et al., 2017). Currently, there is a scientific and technical crisis because of limited empirical data of adequate quality, low funding to improve the situation and development of new technologies, and skilled personnel in measuring erosion rates (Stroosnijder, 2005). The crisis leads to the development of erroneous erosion prediction models since the data are usually taken at inappropriate scales (García-Ruiz et al., 2017). Based on this observation, Boardman (2006) concluded that the soil erosion data in most parts of the world are woefully inadequate, suggesting that little real progress has been achieved despite decades of effort in measuring soil erosion. Additionally, soil formation rates are even less understood in most parts of the world, making soil erosion assessment extremely uncertain (FAO, 2019). Considerable progress has been made toward improving the understanding of soil erosion mechanisms, though grey areas still exist (Li et al., 2018). Areas such as evolution of threads into rills during an erosion event and differentiation of transport and supply-limited removal of coarse and fine material are still unclearly understood (García-Ruiz et al., 2017). Boardman (2006) observed that soil scientists seem to avoid the ambitious questions in favor of more limited, easy-to-answer ones when studying the soil erosion science.

Measuring of soil erosion has been a top target for scientific research and government programs since the beginning of the twentieth century, and due to a variety of reasons still remains a highest research priority (García-Ruiz et al., 2017). Borrelli et al. (2017) emphasized the need to accurately measure the rates of soil erosion and sediment yield at regional scales under present and future climate and land use scenarios. Regrettably, there are high variations in erosion rates, with almost any rate possible irrespective of slope, climate, scale, land use and other environmental characteristics (Baumhardt et al., 2015). Nevertheless, some analyses showed many general features such as positive relationships of erosion rate with slope and annual precipitation, and a significant effect of land use. Agricultural lands have the highest erosion rates where the forest and shrub areas reduced the rates of erosion (Li et al., 2018).

Despite these general trends, there are noted variations that has not been explained, for example by combining the factors, but is related, at least partially, to the experimental conditions (García-Ruiz et al., 2017). Cerdan et al. (2010) showed existence of a negative relationship between erosion rate and size of the study area involved. There were also significant differences associated with measurement methods (Bewket & Teferi, 2009). Direct sediment measurement yielding the lowest erosion rates, and bathymetric, radioisotope and modeling methods yielding the highest rates; and a very important effect of the duration of the experiment (Li et al., 2018). This article analyzed and discussed the conventional and non-conventional-artificial radionuclide techniques for soil erosion assessment using secondary data to summarize the current state of understanding and conclusions in previously published studies. Therefore, this summarized the current state of understanding and compared the effectiveness of different soil erosion assessing methods in different areas.

Conventional methods of assessing soil erosion

In an attempt to assess soil erosion, many conventional methods have been developed over the years though the quantification of soil erosion still remains a challenge (García-Ruiz et al., 2017). Sustainable and effective management strategies are required to assess erosion at local, regional and national scales under different types of activities (Rodrigo Comino et al., 2017). Different countries use different methods leading to the design of numerous risk assessment methodologies across different countries (Fiener & Auerswald, 2016). Timely and accurate estimation of soil erosion loss or evaluation of risk has become imperative for many countries. It is also useful to make estimate of how fast the soil is being eroded before affecting any conservation strategies. Due to the nature of the erosion process, erosion control requires a quantifiable and qualitative evaluation of potential soil erosion on a specific site, and the knowledge of terrain, cropping system, soils and management practices (Borrelli et al., 2017).

Estimation of soil erosion rate encompasses a variety of methods that quantify the material which is removed from the hillslopes, transferred to the channel and transported to other channels, lakes and oceans (Vahid et al., 2021). However, the method used (Borrelli et al., 2017) influence the results, because each method tends to be related to a spatial scale or a range of spatial scales, and consequently each method is selected to measure particular erosion processes. Many methods are used in assessing soil erosion and that include experimental plots of distinct sizes, rainfall simulations, check dams and reservoirs to estimate the accumulated sediment, radioisotope surveys, experimental catchments and models (Tundu et al., 2018; Zhao et al., 2019). Nevertheless, most of these methods have been subject to substantial criticism. Some of the commonly used conventional soil erosion assessment methods include the volumetric, contour plotting, laser scanner and simple mesh-bag (MB) methods. A major limitation of measuring soil erosion by the volumetric methods is that the erosion pins are not sensitive enough for most critical soil erosion measurements (Boardman, 2006). The contour plotting frame work and laser scanner methods were designed to remove some error sources associated with the volumetric methods by increasing the sensitivity in measuring soil erosion. Nevertheless, the contour plotting frame and laser scanner methods are only applicable to limited small (at most few m²) plot sites and are also not very useful in vegetated lands (Cerdan et al., 2010).

The MB field method does not attempt to trap the eroded soil and hence cause little disturbance to the natural runoff pattern of a slope. Unlike the volumetric method, the MB can be used to estimate soil erosion in field with flexible plot sizes and shapes. Another noted advantage of the MB is that it is highly sensitive (detection limit (<0.1 t ha⁻¹ [<0.04 tn ac⁻¹]) and reproducible in field applications (Boardman, 2006). However, the application and interpretation of the MB have not been adequately addressed, for example if the method estimates only the eroded soil that still remains on a slope, how does it account for the eroded soil that has been washed off the slope? (Cerdan et al., 2010). Therefore, doubts are cast on its efficiency in assessing total soil erosion and this has to be attended to before using the MB method. The most commonly used method in quantifying soil erosion is the runoff plot method (Vahid et al., 2020). The runoff plots were extensively used in collecting data for the development of the universal soil loss equation (USLE)/revised universal soil loss equation (RUSLE). The runoff plots are most suitable for ranking relative soil erosion among treatments in standardized plots. Nevertheless, the runoff plots are unsuitable for soil erosion assessment in undisturbed field conditions since their artificial boundaries can alter the natural runoff pattern so may not represent natural runoff and soil erosion conditions.

Notably, the experimental plots were useful in generating information on erosion under various land uses, though their use has recently been criticized as a means of deriving erosion rates (FAO, 2015). For instance, the Soil Loss Estimation Model for Southern Africa (SLEMSA) was developed to estimate soil erosion at a local level (Elwell, 1978). The SLEMSA examines factors involved in the erosion process such as soils, climate, topography, vegetation and human influence and is applied in land use decisions for enhanced productivity. The validity of information obtained from experimental plots has been doubted by many scientists due to disturbance to part of the hillslope caused by artificial drainage boundaries, inadequate representation of natural conditions, the exhaustion of sediment in the mid-term, and the problem of extrapolating the erosion rates obtained to larger areas (Bhattarai & Dutta, 2007; Borrelli et al., 2017). Fiener and Auerswald (2016) noted that small differences in the size of the experimental plots resulted to large differences in runoff and erosion, as a consequence of reduced flow velocities and lower probabilities of infiltration in the smaller plots. Therefore, little credence can be given to erosion rates obtained from closed plots of several m², particularly when the area actually producing the water and sediment is uncertain, and that results subsequently extrapolated to large areas (Mg km² yr⁻¹) are probably unreliable (García-Ruiz et al., 2017; Zhao et al., 2019).

Most of the commonly used conventional methods of soil erosion assessment have severe limitations. The methods are associated with point data (measurement profiles) and do not provide information on spatial distribution of erosion. A major disadvantage of the conventional methods is that they are labor-intensive and require long monitoring periods (Meusburger et al., 2014). One of the problems in attempts to compare erosion rates is the uncertainty created by the use of different erosion measurement methods, to ease this problem, the difference between erosion rate and sediment yield has to be clearly defined before selecting a method to use. FAO (2019) defined erosion rate as the long-term balance between all processes that detach soil material and remove it from a site, and those processes that deliver new material, deposit it at the site. Hence, the erosion rates can be negative (net mass loss) or positive (net mass gain). Sediment yield is the mass that is exported from a given landscape unit, and is always a positive quantity (Lupia-Palmieri, 2004). While some methods such as radioisotope surveys measure true erosion rates, others for example plot or stream sediment monitoring, measure sediment yields. However, it is easy to confuse the two measurements as they are expressed in the same units (mass per unit time, or mass per unit surface and time) (Rodrigo Comino et al., 2017). Most methods can only measure the sediment yield from a closed area but radioisotope surveys provide true erosion rates, including intermediate deposition areas (Evans et al., 2017). In some way to improve in assessing soil erosion rate, radioisotope methods have been developed in soil erosion studies (Navas et al., 2005).

Use of artificial Fallout Radionuclide (FRN) in assessing soil erosion

The increasing negative impacts of soil erosion on agriculture and environment, and need for effective soil conservation measures, drive the need for reliable quantitative data on the extent and actual rates of soil erosion that underpin sustainable soil conservation strategies (IAEA, 2014). The need for new approaches for assessing soil erosion to complement conventional methods has led to the development of methodologies based on the use of FRNs as soil erosion tracers. Artificial ¹³⁷Cs and ^{239+ 240}Pu, natural ²¹⁰Pb fallout and cosmogenic ⁷Be can be useful in assessing the soil erosion. The FRNs can be effective in tracing soil distribution in a landscape and currently offer opportunities for effective assessment of soil erosion and deposition rates (Vahid et al., 2020; Walling et al., 2002). FRNs provide estimates of induced soil erosion rates under different environmental conditions (Mabit et al., 2008). When deposited on the ground, the FRNs bind strongly to fine particles at the surface soil and distributed across the landscape primarily through physical processes (IAEA, 2014). This makes the application of the FRNs in estimating soil erosion possible. Arata et al. (2016) observed a strong correlation between the carbon (C) concentration in topsoil and caesium-137 (¹³⁷Cs); therefore, the inventories of ¹³⁷Cs in the soil can be useful in understanding the movement of soil and associated C on a landform. The binding of the ¹³⁷Cs to the soil C will cause removal of both elements by the physical process of soil erosion. As such the radiotracers provide an effective track of soil and sediment redistribution. However, information on the use of the ¹³⁷Cs in assessing soil erosion is still confusing. Chappell et al. (2011) argued that the loss of ¹³⁷Cs is not necessarily proportional to the loss of soil, suggesting that this method may give misleading data.

Different models of the FRNs vary on their effectiveness in the estimation of soil erosion and deposition rates (Table 1). The average erosion rates estimated with the MODERN under uncultivated soil were lower by 3.1 and 1.9 t ha⁻¹ yr⁻¹ than those obtained by MBM and PDM, respectively. The deposition rate of the MODERN was similar to that of MBM and PDM with standards deviations of 0.8 and 0.6 t ha⁻¹ yr⁻¹ under uncultivated and cultivated soil, respectively (Table 1).

Table 1. Comparison of estimated erosion and deposition rates by different models of FRNs in uncultivated and cultivated soils (¹³⁷Cs)

FRN model	Average erosion rate (t ha^−1 yr^−1)		Deposition rate (t ha^−1 yr^−1)
	Uncultivated soil	Cultivated soil	Uncultivated soil	Cultivated soil
MODERN	7.7 ± 1.2	9.6 ± 1.1	4.6 ± 0.8	5.6 ± 0.6
MBM	10.8 ± 0.9	12.2 ± 1.2	5.5 ± 0.8	6.8 ± 0.6
PDM	9.6 ± 1.2	10.2 ±1.2	6.3 ± 0.8	6.2 ± 0.6

± is the standard deviation (SD). Data derived from Ceaglio et al. (2012)

The application of MODERN, MBM and PDM to the uncultivated and cultivated soils showed their potential to convert FRN inventories into soil redistribution rates for the ¹³⁷Cs and with different land uses (Table 1).

The characteristics of the various FRNs are useful in determining soil erosion in different areas and under different land use scenarios. For example, plutonium 239 + 240 (^{239 + 240}Pu) is a homogeneous fallout with long half-life which allow cost reduction and saves times in measuring. Therefore, the application of the plutonium in investigating soil degradation is more suitable than any other FRNs like the ¹³⁷Cs (Arata et al., 2016). Nevertheless, the use of the ^{239 + 240}Pu in assessing soil erosion is still limited because its conversion inventories into soil erosion rates is still challenging (IAEA, 2014). The conversion models are required in estimating the soil erosion rates in arable lands using the FRNs. The models are needed in accounting the specific depth distribution of FRNs in untilled soils and the soil stratification in order to reflect near-natural conditions (Mandal et al., 2019). The Profile Distribution Model (PDM) and Diffusion and Migration Model (DMM) (Walling et al., 2014) are the most widely used and established models in estimating soil erosion rates in untilled soils (Walling et al., 2014). The PDM and DMM were initially developed to convert ¹³⁷Cs inventories into soil redistribution rates so applying such models to ^{239 + 240}Pu inventories may result to significant bias, since the ¹³⁷Cs and ^{239 + 240}Pu have different depth distribution patterns in soils (Mabit et al., 2008).

The ¹³⁷Cs is abundantly present in the environment because of the fallout after nuclear weapon testing in the 50s and 60s and nuclear accidents (Chappell et al., 2011). The ¹³⁷Cs is used as an environmental marker in assessing soil redistribution because of its strong attraction to clay particles and relatively long half-life. Several researchers found a strong relationship between the ¹³⁷Cs and clay contents; and the proportion of ¹³⁷Cs was increasing with an increase in clay concentration (Chappell et al., 2011; Mabit et al., 2008). Suggesting that the ¹³⁷Cs technology can be a reliable tool to accurately assessing soil erosion and erosion-induced carbon losses (IAEA, 2014). The application of the ¹³⁷Cs technology in measuring soil erosion is simple because estimates can be made by one-time sampling.

In assessing soil erosion, the state and dynamics of SOC should be accurately determined because the SOC is influenced by erosion process and the erosion preferentially removes the low-density organic colloids (Parwada & van Tol, 2018). Basing on that, the methodologies for quantitative assessment of erosion-induced loss of SOC and its impact on crop yield using ¹³⁷Cs were developed (Chappell et al., 2011). The ¹³⁷Cs is preferentially adsorbed to clay particles and OM, which are easily eroded, and this can be source of bias into soil erosion estimates (Stroosnijder, 2005). Therefore, Chappell et al. (2011) argued that ¹³⁷Cs cannot effectively be used to provide information about rates of erosion. Contrarily, recent studies have reaffirmed the usefulness and accuracy of radioisotope methods for soil erosion studies (Mabit et al., 2008; Navas et al., 2005). The soil erosion estimates accuracy (%) was shown to decrease with the eroded phases. Accuracy (%) was higher in slightly to moderately eroded area than in very severely eroded phases (Table 2). In Table 2, the accuracy of the ¹³⁷Cs in estimating soil erosion is shown to be influenced by the intensity of erosion. However, in comparison with the conventional methods, the ¹³⁷Cs was more suitable for calculating soil redistribution with 93–99% accuracy noted in severely to very severely eroded phases (IAEA, 2014). The accuracy varied from 96% to 97% in slightly and moderately eroded phases though a 3–4% overestimation was observed (Table 2).

Table 2. Accuracy percentage calculated for erosion rate using ¹³⁷Cs over traditional methods

Eroded phase	Soil erosion measurement	Erosion rate (Mg ha^−1yr^−1)	Accuracy (%)
Slightly eroded phase	Actual measurement	7.6	-
	^137Cs method average	7.90	100
Moderately eroded phase	Actual measurement	11.6	-
	^137Cs method average	12.10	100
Severely eroded phase	Actual measurement	19.44	-
	^137Cs method average	19.26	99
Very severely eroded phase	Actual measurement	33.60	-
	^137Cs method average	31.46	93

Adapted from Manadal et al. (2019).

The application of ¹³⁷Cs in estimating soil loss requires a reference site that has not been disturbed for a long time (>60 years). The Modelling Deposition and Erosion rates with Radionuclides (MODERN) model is commonly used in quantifying inventories of the ¹³⁷Cs. This model assumes that the depth distribution of the selected FRN is the same at the reference and the sampling sites (Arata et al., 2016). If the soil properties of reference and sampling sites are comparable, the mechanisms influencing the downward diffusion and migration of the radionuclide in the soil should also be similar (Walling et al., 2002). However, it will be difficult to establish a site where no erosion or accumulation has occurred for long period in the real world, to provide a reference site for values of ¹³⁷Cs determined for eroded soils (IAEA, 2014).

The observations in Table 2 confirms that the ¹³⁷Cs technology was a better method for estimating soil erosion even in the very severely intensive croplands (Table 2). Unlike the conventional soil erosion estimation methods, the ¹³⁷Cs technology resulted to more accurate results for all types of erosion, such as erosion due to water, wind and gravity (Arata et al., 2016). In comparison with the USLE, the ¹³⁷Cs-based estimates were observed to provide reliable information for net soil loss rates and the USLE gave overestimated rates (IAEA, 2014). The accuracy has an effective basis for validating the use of ¹³⁷Cs in estimating soil and carbon erosion rates on both cultivated and uncultivated lands (Mandal et al., 2019). Besides the high degree of accuracy, the ¹³⁷Cs technology is faster and low-cost way for assessing historic, comparative and long-term soil and SOC erosion (Walling et al., 2002). Nevertheless, the ³⁷Cs may have shortfall because the gamma-ray counting is expensive and some resource constrained researchers cannot afford but there is significant reduction in labor cost (IAEA, 2014).

The artificial FRNs have a distinct advantage over the conventional soil erosion assessment methods because of their long half-life allowing estimation of soil erosion rates for very long periods possible (Table 3). The artificial man-made ¹³⁷Cs was noted to have a half-life of at least 30.2 years so are applicable in estimating soil erosion for very long periods of time.

Table 3. Comparison of ¹³⁷Cs, ²¹⁰Pb and ⁷Be for assessing soil erosion and redistribution

		Energy(keV)	Time					Measurement		Sediment
FRN	Origin		t1/2	Span	Sampling	Scale	Detector	Lab	In-Situ	dating
^137Cs	Artificial man-made	662	30.2 yr	50 yr (MT)^□	Simple	Plot to large watershed	Normal HPGe γ	Easy	Easy	Feasible
^210Pb	Natural geogenic	46	22.8 yr	100 yr (LT) ^□	Simple	Plot to large watershed	Broad energy range HPGe γ	More difficulty	Limited and unreliable	Feasible
^7Be	Natural cosmogenic	477	53.3 d	≤6 m (ST)*	Fine depth incremental	Local plot to field	Normal HPGe γ	Easy	Longer count time cf. ^137Cs	Feasible

*MT- medium-term; LT- long-term, ST- short-term erosion assessment (Adapted from Mabit et al., 2008).

A noted limitation in using the ¹³⁷Cs was the high spatial heterogeneity of the fallout at the reference sites (Alewell et al., 2019). In situations of high spatial heterogeneity, two long-life isotopes, that is ²³⁹Pu (half-life = 24,110 years) and ²⁴⁰Pu (half-life = 6561 years) (Table 3) can be used (IAEA, 2014). The FRNs were better suited in investigating temporal soil redistribution processes where soil degradation due to agriculture practices affects soil properties and landscape processes than the conventional methods (IAEA, 2014). The FRNs were also effectively used to assessing erosion and deposition patterns in mountain grasslands where the extreme topographic and climatic conditions hinder the application of more conventional techniques (Borrelli et al., 2017).

The ¹³⁷Cs- and ²¹⁰Pb_ex-based methodologies in assessing soil erosion are well researched, but there is still need for researches aiming at optimizing the use of ⁷Be. The obtained information from the ⁷Be evaluations can complement to ¹³⁷Cs and fallout ²¹⁰Pb in the assessment of different soil conservation practices (Mabit et al., 2013). A limitation regarding the use of FRNs as soil tracers lies in the conversion of FRN inventories to quantitative estimates of soil redistribution. The conversion process generated different conversion models that are applied by the scientific community but they differ in their underlying assumptions of soil stratification and descriptions of FRN transport processes (IAEA, 2014). This suggests need for the harmonization of these models.

Soil erosion modeling

Modeling is an involving powerful tool in monitoring and estimating soil erosion. Data generated from the conventional and FRNs soil erosion measuring methods can be used in erosion modeling. Numerous simulation models have been developed and widely used to predict soil erosion rates. Nevertheless, it is noteworthy that models are simply representations of reality (García-Ruiz et al., 2017) and not substitutes for field measurements, so their validation requires high-quality field data (FAO, 2014). Many researchers have criticized the perfunctory use of soil erosion simulation models (Alewell et al., 2019; Bhattarai & Dutta, 2007). Generally, models used to estimate soil erosion under the present and future land cover, climate scenarios or to predict erosion in areas for which little quantitative data are available (Evans et al., 2017). For example, Boix-Fayos et al. (2008) used the WaTEM/SEDEM model in a semi-arid catchment that had six land use scenarios with and without check dams. In this study, Boix-Fayos et al. (2008) highlighted the efficiency of check dams in sediment control projects, and the consequences of increasing land cover density. In a similar study, Martin-Rosales et al. (2003) reflected the importance of models in forecasting reservoir siltation, alluvial plain sedimentation, future water resources and the hydrological and erosive response of catchments to extreme events.

Many soil erosion studies showed the need for major improvements to model performance, to overcome issues of data quality/availability and the inadequate representation of erosion and sediment transport processes (Evans et al., 2017; Martin-Rosales et al., 2003). According to Meusburger et al. (2014) large variations in catchment characteristics and the occurrence of non-linear relationships between sediment yield and environmental variables can reduce the quality of the results from the models. In contrast, Boardman (2006) qualified models as “a good thing”, although several problems reduce their applicability, including complexity, data availability and up-scaling from small to larger scales.

The soil erosion models that are widely used at regional scales include USLE/RUSLE, WEPP, LISEM, EUROSEM, SWAT, SLEMSA, etc., each with its own characteristics and application scopes (Evans & Brazier, 2005). The most widely used models are the Universal Soil Loss Equation (USLE) and Revised Universal Soil Loss Equation (RUSLE) (Bewket & Teferi, 2009). Basically, the USLE predicts the long-term average annual rate of erosion on a field slope based on rainfall pattern, soil type, topography, crop system and management practices (soil erosion factors) (García-Ruiz et al., 2017; Parwada & van Tol, 2018). The USLE methodology has been improved that resulted to a revised version of this model (RUSLE) and further enhanced its capability to predict water erosion (Kouli et al., 2008). The RUSLE is an attempt to improve the capability of USLE in using dynamic hydrological and erosional processes (FAO, 2019). The USLE is simple, the relative ease of obtaining input data and its ability to incorporate the effects of soil conservation practices, which means that it can be used as a management tool. Soil scientist are aiming to develop a new generation of models which may have a stronger physical base and more closely related to the mechanics of the soil erosion process than the USLE (Cerdan et al., 2010).

Recently, there are attempts to combining the USLE model and remote-sensing and Geographical Information System (GIS) techniques. This will make soil erosion estimation and its spatial distribution feasible within reasonable costs and better accuracy in larger areas (Kouli et al., 2008). The innovations in the GIS allow the modeling of complex spatial information. The combined application of GIS and erosion models, for example USLE/RUSLE, was effective in estimating the magnitude and spatial distribution of erosion (Millward & Mersey, 1999). However, these models may have limited application in Africa because they were developed using American soil erosion databases. Making them entirely ineffective in other geographical locations other than the United States of America. In fact, numerous researchers have attempted to determine models which are best-suited to soil erodibility estimation in various areas (Rodrigo Comino et al., 2017; K. L. Zhang et al., 2007). Researches were done to compare these models in specific regions or conditions (e.g. watershed/county-scale studies or rainfall simulation studies) (Wang et al., 2016; K. L. Zhang et al., 2007). Unfortunately, no one model was proved superior as researchers normally only state the model that over or under estimates soil erodibility compared to others. Further, the literature contains some comparisons that are not based on observational data (Kouli et al., 2008).

Currently, the Inventory Method (IM) is the only model which is applied for the conversion of ^{239 + 240}Pu inventories into erosion rates for untilled soils (Lal et al., 2013). The IM assumes that the FRNs depth distribution in the soil is characterized by an exponential function (Arata et al., 2016) but the depth distribution of ^{239 + 240}Pu fallout in soil follows a different pattern. The ^{239 + 240}Pu depth profile soils was observed to follow a polynomial profile with the peak content between 3 and 6 cm soil depth (Alewell et al., 2014). The plutonium is preferentially adsorbed by OM most notably by the humic acids hence retained near the surface (Alewell et al., 2014). The plutonium is found in usually found in reduced concentrations in the top soil and this may be due to plant uptake after fallout deposition (IAEA, 2014). The depth of the soil is inversely related to the long-term rate of erosion that has acted on it on the erosional topography (Arata et al., 2016). The soil thickness was observed to be indirectly related to soil loss, since there is a direct relationship between soil depth and plant growth. Shallow soils with marginal rooting depth are more vulnerable to erosion and landslides (Alewell et al., 2014). Therefore, an accurate quantification of the ^{239 + 240}Pu concentration in the top soil is important as a wrong quantification will result to a biased assessment of soil erosion rates considerably.

Recently, the new conversion model MODERN has been proposed (Arata et al., 2016). The MODERN enables to represent the precise depth distribution of any FRN at a given reference site and most importantly it returns soil erosion and deposition rates in terms of thickness of the soil layer affected by soil redistribution processes (Vahid et al., 2021). In estimating the thickness of soil losses/gains, the MODERN aligns total inventory of the sampling site to the depth profile of the reference site (Walling et al., 2014). This conversion model also takes into consideration of other assumptions that are undertaken in the application of FRNs. The common assumptions are soil erosion tracers, for example the uniform spatial distribution of the local fallout and rapid, strong and non-exchangeable adsorption of FRNs to fine soil particles and soil associated redistribution of FRN through physical processes (Konz et al., 2012). Compared to the Diffusion Migration Model and Mass Balance Model, the MODERN does not require transect sampling approach in which the sampling points need to be located along a transect (Alewell et al., 2014).

Conclusion

Measuring soil redistribution is not an easy task because the soil erosion is caused by a number of processes running at different temporal and spatial scales. Conventional soil erosion assessing methods, such as volumetric, erosion plots, hydrological and geodetic, are used for different erosion processes and they cover different spatial and temporal scales but have numerous limitations on the accuracy of the data gathered. Application of isotope tracers can avert the deficiencies of the conventional methods as some naturally occurring radionuclides and stable isotopes in the environment can serve as tracers hence facilitating the investigation of the erosion processes. Generally, the FRN showed to provide a better assessment of soil erosion compared to the conventional methods, however where the resources permit, combining the two may yield best results. The choice of erosion measurement or estimation method depends on the type of erosion, topography, vegetation and budget.

Recommendation

It is difficult to recommend a one-take for all soil erosion assessment method as the suitability varies from place to place. The time period in which the soil erosion data are required should also be considered when selecting on the assessment method to be used. Generally, there is a need for developing new approaches and paradigms in estimating soil erosion.

Acknowledgements

The authors gratefully acknowledge the Zimbabwe Open University for time and resource support received for the study as well as the Marondera University of Agricultural Sciences for the support given to the first and last authors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

Additional information

Notes on contributors

Cosmas Parwada

Cosmas Parwada is an Agriculture and Rural Development Scientist at the Tugwi Mukosi Multidisciplinary Research Institute (TMMRI), Midlands State University, Zimbabwe. He is a soil scientist with a sound background in agronomy. He has published many articles related to crop production and soil fertility and conservation.

Justin Chipomho

Justin Chipomho is a seasoned lecturer and head of Crop Science department at the Marondera University of Agricultural Sciences and Technology, Zimbabwe. He is a renowned researcher in crop production and soil fertility.

Handsen Tibugari

Handsen Tibugari is a lecturer at the Gwanda State University, Zimbabwe. He has extensively researched on potential utilization of sorghum allelopathy for weed management. He has vastly published articles related to crop production and crop protection.