1 Introduction

Soil erosion represents a hazard to food and water security (Batista et al. 2019), it reduces soil depth, increasing soil loss and decreasing the capacity of soils to storage water (Li et al. 2009). It has a significant impact on the productivity of cultivated lands, affecting physical, chemical and biological functions of soils, increasing the carriage of nutrients and a wide variety of other chemical substances, with the consequently eutrophication of water bodies (Delgado et al. 2018). At the same time, sediment yield may affect adversely various parameters at a watershed, as for example causing the decrease of the storage capacity of reservoirs. Several of these problems might be directly related with sediments transported downstream in the water flow. In this context, nonstop soil loss will probably turn out to be a major problem for global conventional agriculture (Montgomery 2007).

Furthermore, soil erosion is considered to involve sediment detachment from soil surface, including the impact of raindrop and also the flowing water (Jain and Kothyari 2000). Erosion processes are toughly affected by many factors, besides rainfall (Fang et al. 2013). In areas with changing land-use processes interactions between soil, land-use, sediment production and sediment transport can be quite complex (Tramblay et al. 2010). In addition to geological erosion, human practices such as forest clearing, agriculture, livestock farming, mining and urbanization might induce a faster and more detrimental process of erosion (El-Swafy et al. 1982). Land-use changes might increase sensitivity of landscapes to disturbance, making them more vulnerable to extreme events (Boardman 2006). Other land-uses, as for example urbanization -due to its impervious or semi-impervious effect-, contribute to generate more runoff and increase peak flows, carrying out the consequently growth in the frequency of floods, erosion processes and also, the transport of some pollutants (Pappas et al. 2008).

Moreover, understanding how erosion affects the productivity of soils under different farming conditions should be studied and well understood in order to be able to develop lasting management strategies (Li et al. 2013). Among the available soil erosion and sediment yield models for assessing water erosion worldwide, the Universal Soil Loss Equation (USLE) created by Wischmeier and Smith (1978) and its modified and revised versions are some of the most used models (Sadeghi 2004; Sinshaw et al. 2021). The USLE model predicts soil erosion in order to contribute to conservation planning (USDA 1997). The model concept is not based on process description and simulation but rather on understanding a process, capturing the measureable parameters and delineating a mathematical algorithm (Alewell et al. 2019).

Considering areas with data scarcity, the USLE and its revised versions appears to be suitable for reaching a preliminary estimation of soil erosion (Delgado 2018; Kamamia et al. 2022). In this context, besides the need of additional features such as various data sources, data scaling, and higher data integration and algorithm complexity (Sinshaw et al. 2021), the analysis of satellite data within a Geographic Information System (GIS) environ appears to fit for preliminary research (Ali and Hagos 2016). For example, at a regional scale Borrelli et al. (2013) estimated that in South America, during the period from 2001 to 2012, an 8% increase in soil erosion occurred as a result associated to land-use change, mainly related to deforestation and the expansion of cultivated areas in Argentina (41.6%), Brazil (19.8%), Bolivia (37.8%) and Peru (5.9%).

In particular, proper indicators for assessing environmental changes are infrequent in Argentina (Viglizzo et al. 2003). Although there are specific scientific studies regarding erosion and soil loss at national and regional scales (Irurtia et al. 2007; Gaitán et al. 2017; Casas et al. 2022; Fernández et al. 2022), data available is generally scarce at a local scale for plain territories. Most of the time this issue is primarily associated to the absence of hydro-environmental monitoring networks (Cesanelli and Guarracino 2011), and therefore there is a limited data source for many environmental studies.

For that reason, the aim of this work was to stablish the current soil loss in a periurban flatland watershed of the Pampean region, named as El Pescado Creek, located in the NorthEastern of the Buenos Aires province (Argentina), and compare its value with past and future scenarios. The future scenario was established according to local land-use estimations (considering local trends and accounting with delimited areas derived from different buffer areas) together with global projections. Considering that Delgado et al. (2020) already showed that the watershed of El Pescado Creek is experiencing an intense land-use change process, we trust that the results of this research will help to elaborate management strategies for this territory, providing useful information for local planning. Only one publication addressed this subject in El Pescado Creek (Besteiro and Gaspari 2012), although only considered a reduced area of the watershed. As mentioned by Alewell et al. (2019), soil erosion modelling is essentially needed on both small and large spatial scales for planning, management and policy measures. Thus, in order to provide further knowledge on this matter and based on the limited existence of previous research about this topic in this flatland territory -where slope is not an issue- we believe that this research will probably help to recognize the way other variables act regarding soil loss.

The novelty of this research relies on the fact that it represents a first step in the understanding of soil erosion in this flatland watershed, being able to provide useful insights for managing socio-economic growth in this changing territory. At the same time, this methodology could be extrapolated to many other watersheds of the Pampean region, which are going through similar land-use and land-cover change process, largely towards agriculture and unplanned urbanization. Up to now, there is no previous research regarding this issue at the whole watershed scale of El Pescado Creek. Certainly, these findings will contribute to a further comprehension of the relationship between soil loss and anthropic and environmental factors, which may vary over time and space, and consequently might generate an increase/decrease in soil loss rate for a certain region. Additionally, we consider that this knowledge is able to be used by policy-makers in the establishment of a future territorial planning.

2 Materials and methods

2.1 Study area

The Pampean region, placed in Center-Eastern Argentina, comprises an extensive area of 613,532 km2, with plentiful water resources and a mainly plain topography (Aliaga et al. 2021). The watershed of El Pescado Creek, with a total area of 349 km2, is placed in the NorthEastern of the Buenos Aires province, draining its water into the Río de la Plata (Fig. 1). The climate is humid-temperate in this area of the Pampean region, the mean annual precipitation is around 1083 mm for the period 1986–2019. According to the annual water balance, there is a water excess from February to November and a water deficit in the summer season, between December and January (Moreiro Varela et al. 2020). It has been reported an increase in the potential erosiveness of rainfall for the period 1961–2007 (Besteiro and Delgado 2011).

Fig. 1
figure 1

Location map of El Pescado watershed, in the Pampean region and the Buenos Aires province (Argentina). Main land-uses are mapped for the watershed, with detailed captures of satellite images taken from Google Earth and pictures taken in field surveys carried out by the research group

Full size image

It is considered that groundwater is stored in a multilayer aquifer, composed of an unconfined and a semi-confined aquifer, detached by a silty clay layer. The unconfined aquifer is directly recharged by infiltrated rainfall, being the semi-confined aquifer indirectly recharged from it (Kruse et al. 2013). Regarding topography, the elevation is below 30 m asl (meters above sea level), with a north-eastward direction for regional groundwater flow. The hydraulic gradients vary from 1.8 to 0.5 m/km (Carol et al. 2012). Based on the National Institute of Agricultural Technology (Instituto Nacional de Tecnología Agropecuaria (INTA)), a major area of the watershed is covered by Aquic Paleudoll, with a significant presence of Typic Paleudoll in the main watershed´s divide. Typic Natracuert and Natracualf are found in the form of disperse patches. Soils are characterized as Miscellaneous in most of the area near the sink. In the floodplain, soils are described as an Undifferentiated complex (Delgado et al. 2022).

The province of Buenos Aires is characterized as the one with the highest demographic and industrial concentration in Argentina, together with the leading agriculture and livestock production (Rodriguez Capítulo et al. 2010). The watershed of El Pescado Creek was declared as a protected landscape (Law 12,247 of the Province of Buenos Aires 1999), although no guidelines on land-use and land-coverage have been established. This lack of planning contributes to an intensification of agriculture activities and unplanned real estate developments. These trends have led to modifications in the natural relationship between infiltration and surface runoff (Delgado et al. 2020) and consequently in the water cycle. Most of the upper watershed of El Pescado Creek belongs to the Horticultural Belt of La Plata HBLP (locally identified as the Cinturón Hortícola Platense), which is known as one of the largest productive regions of fresh vegetables in Argentina. Also, some few semi-urbanized areas are rising inside the watershed (mainly closed to the northern and north-western watershed divide), thus, in this areas previous extensive livestock fields have become impervious or semi-impervious lands. Grasslands, principally composed of cespitose grasses, characterizes the original vegetation of the region. Also, some native forest species can still be found, such as Parkinsonia aculeate (Cina cina) and Celtis tala (Tala). The extensive livestock raised in this territory is primarily represented by bovines. The exotic forest species are represented by eucalyptus and pines.

2.2 Soil loss equation within a geographic information system environ

Estimations were done for 1985 and 2021, in order to compare soil loss from two different land-use scenarios. As mentioned by Hang et al. (1995), in the early 1990s, intensive agronomic technologies, such as greenhouses, were introduced in the country, in areas with horticulture-production systems as the one placed in La Plata. Thus, the time period selected aimed to consider two dissimilar circumstances, before (1985) and after (2021) the introduction of intensive agronomic technology in the area. The year 1985 was chosen based on the accessibility to satellite images, that enabled to establishing land-use and land-cover, and consequently the cartographic representation of the C and the P factor. The spread of agriculture (with and without greenhouses) turned out as a result of the mentioned introduction of intensive agronomic technologies.

The open source GIS QGIS was used for geospatial data analysis. The Universal Soil Loss Equation (USLE) proposed by Wischmeier and Smith (1978) was used to estimate soil erosion (Eq. 1). Inputs are based on information related to precipitation, soil characteristics, percentage and length of slopes, and land use characteristics together with soil conservation practices.

$$A=R*K*LS*C*P$$
(1)

where A: average soil loss per unit of area (Mg ha−1 y−1), R: rainfall-runoff erosivity factor (M J mm ha−1 h−1 y−1), K: soil erodibility factor (Mg ha h MJ−1 ha−1 cm−1), LS: topographic factor (dimensionless), C: cover management factor (dimensionless), P: support practice factor (dimensionless).

2.2.1 Description of the five factors

2.2.1.1 Rainfall-runoff erosivity factor (R)

It reflects the rainfall energy that may generate soil erosion. It is calculated as a function of the total energy of the event and its maximum intensity in 30 min. Since available local data only refers to daily statistics, it was not possible to calculate a current R factor for this watershed following the original methodology, thus, the R factor was established considering global available data elaborated by Panagos et al. (2017). This global erosivity map was established at 30 arc-seconds (~ 1 km), based on a Gaussian Process Regression (GPR), and can potentially cover gaps, where erosivity has not been estimated due to lack of available data. Its generation involved the collection of high-temporal resolution rainfall data, calculation of the R factor for each rainfall station, normalization of calculated values with rainfall data, and finally, spatial interpolation of the factor. Units were adjusted before applying Eq. 1.

2.2.1.2 Soil erodibility factor (K)

This factor represents a mean value of the total soil response to hydrologic processes and surface erosion. It reflects the complex and combined action of different soil properties such as permeability, organic matter, texture and structure. According to Shabani et al. (2014), it is difficult, and often impossible, to set up the standard plot for its determination in every soil type.

Data was acquired from INTA and CISAUA (2006) soil maps. The K factor was estimated for each soil series based on Eq. 2 (Wischmeier and Smith 1978), and then according to the percentage of incidence, the mean pondered value for each soil subgroup was calculated.

$$100*K=2.1*{M}^{1.14}*\left({10}^{-4}\right)*\left(12-a\right)+3.25*\left(b-2\right)*\left(c-3\right)$$
(2)

where, M refers to the size of particles, defined as the percent of silt (0.1–0.002 mm) multiplied by the quantity 100, minus the percentage of clay, a = organic matter (%), b = soil-structure code, c = profile-permeability class.

2.2.1.3 Topographic factor (LS)

This factor considers the slope length (L) together with the slope steepness (S). It is defined as the relationship between soil loss for a certain length and soil loss for a 22.1 m plot, (considering the same soil type). Steep slopes correspond with higher values of LS factor, flatland areas tend to have smaller LS factors, which consequently decrease soil erosion.

It was automatically defined within QGIS, by the process r.watershed, using as the input file the Digital Elevation Model provided by the National Geographic Institute of Argentina (Instituto Geográfico Nacional- IGN) which brings combined data obtained by the SRTM and ALOS satellite missions. The scenes 3357-19, 20, 25 and 26 were downloaded and joined in a mosaic.

2.2.1.4 Management and cover factor (C)

This factor considers the soil protection brought by the vegetation cover and represents all the combined result of the interconnected cover and management characteristics. It expresses the relationship between soil loss in a cultivated plot and in a plot without vegetation cover, with all the other conditions undisturbed.

For the current land-use scenario the C factor was determined for year 2021, land-use and land-cover was manually defined using images provided by Google Earth© together with gathered global annual land-use and land-cover data (Karra et al. 2021) provided by ESRI at 10 m resolution.

2.2.1.5 Support practice factor (P)

This factor comprises practices that affect soil erosion by transforming flow direction together with the volume and rate of surface runoff. It involves the relationship between soil loss in a cultivated plot with a conservation practice and soil loss generated in a cultivated plot without any conservation practice. This factor expresses the effect of conservation practices such as contouring and terracing, among other, on soil erosion (Devatha et al. 2015).

Estimation of soil erosion for the past scenario (1985) consisted on modifying the C factor map, according to established land-use for that year. Other factors remained unchanged in this estimation. In the case of the future scenario (2050), R and C factors were changed. Regarding the R factor, we used the 4.5 projection for 2050 elaborated by Panagos (2022), which constitutes the intermediate simulation of what R factor will probably be in future years. The C factor was established according to land-use estimations for 2050, considering the already ongoing expansion of Crops, Urbanization and Greenhouses. Buffers were applied in order to increase the area of occupancy of these land-uses, providing special priority to Urbanization among any other land-use. Based on the fact that agriculture (represented by Crops and Greenhouses land-uses categories in this research) has already speedily spread in the watershed, we assigned an 100 m buffer distance for 2050 scenario, and increased to a 500 m buffer for Urbanization land-use based on growing real estate pressure in the area, representing the category with the major increment regarding land-use projection for year 2050. We based this decision in the fact that agriculture has already reached most of the surface available for cultivation in the watershed, as it can be seen in Fig. 1. Also, the need of land for housing construction is dramatically increasing around the city of La Plata, pushing people to establish in peri-urban and rural areas, with no guideline or planning, and often without basic services such as water supply or sewage services.

The USLE equation was calculated by multiplying the previously five mentioned factors with each other. Thus, annual soil loss rate (Mg ha−1 y−1) was obtained for the years 1985, 2021 and 2050. Each soil loss map, expressed in Mg ha−1 y−1, was later classified (within the command Reclassify in QGIS) into categories of risk of soil erosion, according to the classifications established by FAO (1980) and Teng et al. (2019). In this way, every pixel inside the watershed of El Pescado Creek was assigned to a specific category. Ranges established by FAO (1980) for soil erosion risk were: low (< 10 Mg ha−1 y−1), moderate (10–50 Mg ha−1 y−1), high (50–200 Mg ha−1 y−1) and very high (> 200 Mg ha−1 y−1). Classification done by Teng (2019) established the next erosion grade categories according to soil loss rate: limited (< 0.01 Mg ha−1 y−1), slight (0.01–1 Mg ha−1 y−1), light (1–2 Mg ha−1 y−1), moderate (2–11 Mg ha−1 y−1), intense (11–20 Mg ha−1 y−1), extremely intense (20–30 Mg ha−1 y−1) and severe (> 30 Mg ha−1 y−1).

3 Results

3.1 Individual factors of USLE

Geo-spatial distribution of each factor for year 2021 (current scenario) is shown (Fig. 2). R factor mean value in the watershed was 233.74 ± 1.14 M J mm ha−1 h−1 y−1, it expressed its minimum value (230.05 M J mm ha−1 h−1 y−1) mainly in the headwater area, and its maximum value (237.30 M J mm ha−1 h−1 y−1) in the middle of the watershed (Fig. 2a). The calculated K factor (Table 1) expressed a mean value of 0.396 ± 0.13 Mg ha h MJ−1 ha−1 cm−1 for the watershed, meanwhile the minimum and maximum values assigned to the soil subgroups were 0.15–0.711 (Mg ha h MJ−1 ha−1 cm−1), respectively. The minimum values were located in the flood plain of El Pescado Creek, increasing towards the main water divide of the watershed, and reaching its maximum value in the sink (Fig. 2b). As it is expected, being the USLE model a multiplicative equation, a higher value of a factor will be corresponded to a higher value of soil loss. Regarding topography, the LS factor mean value in the watershed was 0.08 ± 0.20, with the minimum and maximum value of 0.03 and 4.99, respectively; this factor is homogenously distributed in the watershed, with rather small numeric values derived from the scarce slopes in the area (Fig. 2c). The C factor expressed a mean value of 0.19 ± 0.07, with a minimum and maximum value of 0.013 and 0.30, respectively. This factor showed a scattered distribution of its minimum values through the whole watershed (associated to Forest land-use), intermediate values were mainly found it the lower and middle watershed (Grassland), meanwhile maximum values were preferably associated to the upper watershed (Crops, Urbanization and Greenhouses) (Fig. 2d). The values of C factor (Table 2) were assigned following previous research (Wischmeier and Smith 1978; Hill et al. 2015; Loaiza and Da Silva 2018). Since no conservation measures were found in the watershed, the assigned P factor was equal to one (1), consequently, it was not represented in Fig. 2.

Fig. 2
figure 2

Maps representing the USLE factors for current scenario. a Rainfall-runoff erosivity factor (R), b Soil erodibility factor (K), c Topographic factor (LS) and d Management and cover factor (C). The P factor is not represented since its value is 1

Full size image
Table 1 Estimated values of K factor (Mg ha h.MJ−1 ha−1 cm−1) for each soil subgroup, and its area in the watershed
Full size table
Table 2 Assigned values of C factor (based on bibliography), and its area
Full size table

3.2 Annual soil loss

The results obtained in the USLE model (Mg ha−1 y−1) for the three scenarios were classified in grades of soil loss following FAO (1980) and Teng et al. (2019) classifications (Fig. 3). For the past scenario (1985), according to FAO (1980) the 99.3% belonged to the category of Low erosion risk and only the 0.7% corresponded to the Moderate category (Fig. 3a). Using the Teng et al. (2019) classification (Fig. 3b), the most relevant incidence was estimated for the categories of Slight erosion grade (77.9%), Light (14.1%) and Moderate (6.9%).

Fig. 3
figure 3

Classified USLE value in the watershed of El Pescado Creek. According to FAO (1980): a (1985), c (2021) and e (2050), and Teng et al. (2019) classifications: b (1985), d (2021) and f (2050)

Full size image

For the current scenario (2021), most of the watershed (98.3% of the area) belongs to the category of Low erosion risk, which is represented by values < 10 Mg ha−1 y−1, following the FAO (1980) classification (Fig. 3c). When Teng et al. (2019) classification was applied, most of the watershed expressed a Slight erosion grade (69.5%), together with Light (10.9%) and Moderate (17.2%) grades (Fig. 3d).

Finally, for the future scenario (2050), the 95.1% of the watershed belonged to the category of Low erosion risk using the FAO (1980) classification, and the 4.4% and 0.5% to the Moderate and High category, respectively (Fig. 3e). But when the Teng et al. (2019) classification was applied (Fig. 3f), the 23.4% corresponded to the category of Slight erosion grade, and the 45.0% and 27.2% to the Light and Moderate categories, respectively. In this scenario, appears the occupancy of the higher categories of erosion grade (which had no incidence in the past and current scenarios), representing: 2.4% in the Intense category, 0.8% and 1.1% in the Extremely intense and Severe categories, respectively.

4 Discussion

The El Pescado Creek watershed showed a remarkable growth of agriculture and urbanization in the last thirty years, and this trend is estimated to be continued in the future, certainly, without any planning or regulation taken into account. It is known that only reaching the integration of various processes of the coupled human-natural system would be possible to understand the effects of human society and economic activities on the environmental system (Liang and Zhong 2023). In their global scale study Beyene and Youssef (2023), highlighted the mechanisms through which economic activities might be influencing the environmental health and ecosystem vitality, and suggested the need for policy responses to manage these impacts, and also for reaching a more holistic understanding of these complex interactions.

In a global study regarding soil loss using the modeling platform Global Soil Erosion Modeling (GloSEM), Borrelli et al. (2020) proposed that water erosion is a common phenomenon under all climatic conditions all over the world. At the same time, socio-economic development of drainage watersheds has had a relevant and evident impact on the soil loss rates in every landscape (Poesen 2018). Although this tendency is described at a global scale, the level of incidence of different variables may vary between different types of watersheds according to landscape characteristics. For example, Rajbanshi et al. (2023) after their meta-analysis of almost a hundred research manuscripts stated that the outcomes of different water and soil conservation practices might vary greatly across different natural set up and management scenarios. The same authors also mentioned that some conservation practices, such as mulching and cover crop, are less suitable in the long term, even though they are highly efficient in reducing soil loss and surface runoff at the initial stage of establishment.

It is to be noticed that although the western area of the watershed of El Pescado Creek is affected by the major land-use and land-cover change process, soil erosion rates remain low and similar to other areas of the watershed (as for example the eastern area), where current and projected anthropic use is quite smaller (Fig. 3). In this regards, authors such as Milazzo et al. (2022) showed how grassland plays a relevant role on the regional cumulative erosion mitigation, as the grassland total conversion to permanent cropland can raise the regional cumulative erosion around 15%. In this context, it is evident the incidence of the scarce slope, reflected in the LS factor, in order to decrease soil loss. Slight topographic differences between the interfluvial sectors and the floodplain next to the watercourse determine the existence of very slight slopes along the watershed, contributing to minimize soil loss (Guerra et al. 2017; Boardman et al. 2022). The El Pescado Creek watershed has a very low LS factor, with an already mentioned mean value of 0.08 (± 0.20), which undoubtedly contributed to diminish the calculated values of soil erosion. In addition, it is noteworthy that if this turns out to be the determinant factor in the erosion process in this study, it participates both at local and regional scale, since the slopes in the studied watershed have a regional representativeness, characterizing most of the drainage watersheds of the Pampean plains (Graziano et al. 2021). Perhaps, using a finer resolution DEM might have improved the accuracy of the LS factor, although requiring a higher computational demand of the analysis, with more processing time and storage space (Makhdumi et al. 2023).

The analysis of the remaining factors showed that some of them -obtained by global estimates- may not adjust to a local watershed scale. In the case of the R factor, due to the significant amount of data required for the USLE model calculation of rainfall erosivity, studies based on scarce accurate data have used various equations depending on the temporal resolution and accessibility of the rainfall data (Makhdumi et al. 2023). In our study, although there are data done at national scale (Rojas and Conde 1985; Gaitán et al. 2017), we used global values (Panagos et al. 2017) in order to be able to compare with the 2050 projection done by Panagos et al. (2022), and also, because the global geospatial distribution data were the only available one to be downloaded in a GIS format file. Thus, there was no access to any other reliable R factor determination for this geographical area.

Another limitation to the model in local context might be the use of outdated information that could be indirectly affecting the factor´s value. Such is the case of the K factor, which was calculated using the official Soil Series descriptions (mostly based on field data obtained in the early 1970s), thus, some parameters as the values of organic matter might have probably changed over time. This fact will certainly have a direct impact on soil loss rate, since organic matter is directly involved in the K factor, as showed in Eq. 2. Also, as mentioned by Shabani et al. (2014), slope may be playing a key role in the K factor, which is markedly influenced by land-use and permeability, as well as by the content of lime, sand, silt and organic matter. These authors mentioned that although the effect of slope is already considered in the USLE model (through the LS factor), its incorporation in estimating the K factor has been found to be useful. Our research does not deal with this matter, but it might be an interesting approach to explore.

A similar situation occurs regarding the C factor, which could be adjusted to a more detailed discrimination within a same category of land-use and land-cover, as for example, in the case of the Crops (splitting between horticulture, soybean, corn, sorghum or other cultivated species in the region), or Grassland, which includes different types of extensive livestock in the watershed (ovine-bovine-equine). Further research could follow this way, in order to be able to study each specific effect on soil erosion, perhaps it would be useful to consider dividing the watershed in sub-watersheds, in order to reach a more in detail research. These problems that arise when defining the factors to be used in erosion models are faced by most modeling studies where there is an understanding gap about the trustworthiness of the obtained results (Batista et al. 2019). Despite this, the generation of this type of model is still a relevant tool in soil erosion assessment (Borrelli et al. 2022), since these methodologies own a relevant role in supporting policy assessments of decision-makers. In this context, we underline that the generated model symbolizes a first stage regarding the research of soil erosion and its driving forces in El Pescado Creek watershed. Through this investigation we estimated an increase in soil loss from the past scenario -1985 (mean of 0.85 Mg ha−1 y−1) to the current one -2021 (mean of 1.52 Mg ha−1 y−1), with a trend that will continue towards the future scenario -2050 (mean of 3.07 Mg ha−1 y−1). Our results are in accordance with those obtained by Borrelli et al. (2013), enabling to estimate a mean value of 1.28 Mg ha−1 y−1 at this watershed scale, for year 2001. We highlight that these results indicate erosion rate at regional scale, considering that the minimum values of the LS factor (similar in mostly all the flatland drainage watersheds of the Pampean region) are determinant in present results.

As already mentioned, soil erosion is driven by the complex interaction between topography, climate, and land-use, thus, changes in land-use and land management might generate highly non-linear effects. Understanding how soil loss might be minimized by natural landscapes is therefore a major priority for landscape managers (Hamel et al. 2015). In order to validate soil erosion models and expand understanding of the process, we must conduct well-planned on-site measurements at both the local and regional scales (Makhdumi et al. 2023). Considering that soil erosion causes both on-site degradation of soil properties and productivity and off-site problems (Gashaw et al. 2021), further studies should include field campaigns pursuing to monitor sediment yield. Lacking annual determinations (as the USLE model deals about annual estimations), sediment yield for a single rainfall event could be measured and also modeled throughout other models, such as the Modified Universal Soil Loss Equation (MUSLE) which is based on the USLE methodology, but adjusted to a rainfall event. Although ongoing monitoring and assessment to track changes in soil erosion rates over time is relevant, it would be very hard to maintain, since vandalism is frequent in the area. At the same time, when an erosion plot is installed in a private field, most of the maintenance work and agronomic practices will rely on the local producer. Thus, in this type of watershed, it would be surely very hard to install a soil loss plot and get reliable data from it over time. Another possibility in order to cope with this issue could be to establish indoor simulations using a laboratory rainfall simulator, but, undoubtedly, it would not be able to fully represent the real field conditions, such as slope length and in depth root system, among others.

Although we are aware that field assays and monitoring networks will be needed in order to validate our results in further studies, both in plot and in watershed scale, we consider that the USLE model was properly used to estimate soil loss in El Pescado Creek watershed, which is an area with scarce field available records. Its simplicity makes it suitable for initial geo-spatial studies (Alewell et al. 2019), similar methodologies could be also implemented in nearby watersheds in order to reach future territorial management guidelines for this region of Central Eastern Argentina. This information may contribute to a deeper comprehension of the possible soil loss scenarios and their interactions with anthropic modifications. The ability to effectively adapt to climate change depends, among other, on the amount of understanding, funding and chances available to farmers in order to address to that modification (Nazari et al. 2015). Moreover, the role of government, policymakers, and stakeholders is essential in formulating policies and strategies to foster socio-economic growth in a sustainable way (Mukherjee et al. 2023). Achieving sustainability through land systems is challenging because multiple beliefs and values exist, and also because benefits from land are unequally distributed; and because occasionally actors mobilize multiple, sometimes conflicting, visions of justice (Meyfroidt et al. 2022). In the watershed of El Pescado Creek, as it has already been mentioned, there is a wide variety of interests, it represents a complex territory not only for the land-use and land-cover change process taking place. Agriculture is spreading over a great area, but also unplanned urbanization, these two land-uses might carry out diverse additional effects, but, at the same time, this watershed constitute a declared protected landscape. Thus, we believe that on a first step, the already existing regulations and laws regarding allowed and not allowed activities must be taken into account.

In order to diminish possible negative effects, policymakers, stakeholders, and local farmers together with the support of the academic sector, should try to find a way to reach sustainable intensification. This concept, as described by Pretty (2018), seeks to improve synergies between landscape components and agriculture, being a main concern for the Sustainable Development Goals of the United Nations (SDGs). We could say that our research deals only with Goals 11 (Sustainable cities and communities) and 12 (Responsible consumption and production), but, indeed, it also reaches other Goals, such as 6 (Clean water and sanitation) and 15 (Life on land). As mentioned by Meyfroidt et al. (2022) these universal goals are crucial to monitor efforts toward sustainability, but, at the same time, it has to be considered that appropriate solutions for a given context might not be useful in other contexts or fail to achieve the balance of benefits desired by stakeholders. Consequently, different governance involvements at multiple scales from local to global are needed.

Nowadays it is well known that the transition to multifunctional landscapes is a process with major potential to enhance biodiversity and nature's contributions to people, providing at the same time more nutritious food and stable crop yields (Garibaldi et al. 2023). It must be taken into account that understanding the perspectives of diverse stakeholders is a major challenge for ecosystem management, which is much more complex than setting scientific limits on resource use (DeFries and Nagendra 2017). In this way, and in accordance with Pretty (2018), we consider that agroecosystems will need to adapt in order to be able to generate an extended source of ecosystem services -defined by Díaz et al. (2015) as the benefits that people obtain from ecosystems- including the production of food together with water and soil conservation, among others.

Although agriculture is essential for development of human population, the lack of planning, management and control measures might carry negative consequences to the ecosystem, such as fertilizers and pesticides surplus, known as none point source pollutants (Davie 2008). At the same time, urbanization is increasing in El Pescado Creek and neighbor watersheds, with apparently no guidelines, although this watershed represents a protected landscape at the province of Buenos Aires (Delgado et al. 2020). Allen (2003) have already explored the main features that make environmental planning a distinctive process, defining the peri-urban interface as an “uneasy” phenomenon, generally characterized by the loss of “rural” aspects (agricultural land, natural landscape) or the lack of “urban” attributes (low density, lack of accessibility, lack of services and infrastructure). This multiplicity of uses and conflicts create new livelihoods and new social relations, generating a heterogeneous mosaic of natural ecosystems and productive, agricultural and urban systems. Also, the planning of this territory must meet multiple aspects: quality of the landscape, coexistence of diverse ecosystems, conservation of different resources, economic development and social tensions, and others (López-Goyburu and García-Montero 2018). Reaching a comprehensive analysis of risk and vulnerability at the local scale might help in the development of specific actions to manage vulnerability and reduce the risk of a variety of hazards. Involvement by the population in the generation of local efforts favors the development of strategies that could improve urban planning and increase social resilience (Palacio-Aponte et al. 2022). There is a growing consensus about policies addressing urban sprawl containment not only contributing towards optimization of the use of infrastructure and reduction of the cost of providing public services, but also mitigating environmental impacts related to landscape degradation and climate change, closely related with ecosystem functions, hydrological processes, and urban quality of life. Undoubtedly, this policy depends on the confluence of many factors and forces concerning territorial governance, in order to guide and regulate urban development to more sustainable forms (Menzori et al. 2021).

At recent studies at global scale, land-use change has been identified as the most important driver of biodiversity loss for many well-studied taxa (amphibians, mammals, reptiles, birds). Also experts reported that biodiversity loss is driven indirectly by production and consumption, human population dynamics, governance, trade, and technology (Isbell et al. 2023). Sustainable development and sustainability represent challenging problems due to their socio-economic and environmental dimensions, considering also the many different ways of management, legislation, governance frameworks at local, national or international scale, and their possible combinations (Teodosiu et al. 2022). In this context, and pursuing the sustainability of the watershed, is that we consider that territorial planning must be incorporated into local policy maker’s agenda immediately. In this regards, Taoumi and Lahrech (2023) suggested a new academic word, the so called “socio-eco-effectiveness”, and expressed the need to establish a multidisciplinary model that addresses the social aspect as a variable with the economic and the environmental ones, in the objective functions, rather than considering it as a constraint. This could help policymakers to support food security and to promote agricultural production within the SDGs directives around the world.

5 Conclusions

This study case showed that erosion might be an issue in spite of dealing with flatland territories; although the incidence of the scarce slope is reflected in the LS factor -which contributes to minimize soil loss- erosion still could be found in the watershed. In flatland areas water movement, after rainfall reaches the soil surface, tends to be vertical (infiltration and evapotranspiration), instead of horizontal (runoff) usually seen in hilly or mountainous areas. Hence, it might be predicted that almost no soil loss would be found in this region, but instead, we highlighted that there is a soil erosion process taking place already, and that indeed it might increase towards future scenario.

These findings showed the need of institute sustainable management strategies in the watershed, as a first step towards territorial management, in order to diminish current soil erosion and also to avoid its potential increase over time. Although current soil erosion rate is still low, land-use change process towards agricultural and urbanized areas remains under no guidelines. We were able to identify soil loss rates modifications during past-current-future scenarios. The most relevant land-use and land-cover changes were related with the growth of agriculture -characterized by greenhouses setting up in the upper watershed, together with different crops scattered spread in the rest of the watershed-, and the expansion of urbanized areas.

This first step regarding the understanding of soil erosion in this watershed could also be replicated in nearby watersheds with similar characteristics, since there are many other watersheds with similar natural features that are going through a comparable land-use and land-cover change process, keeping equivalent edaphic, climatologic and topographic characteristics. Perhaps the changes are not exactly in the same way, but, as already mentioned, the regional tendency is moving towards the replacement of grassland by agriculture and urbanization, most of the time without any planning or guidelines. A deep understanding of the effects of individual factors, together with the multiple variables affecting them, is needed. This information may contribute to a further comprehension of the possible soil loss scenarios and their interactions with anthropic modifications, considering that environmental, social and economic conditions might generate undesirable stresses, thus, affecting not only a production system itself, but also the people involved, as for example, farmers. Soil loss might generate several consequences, not only in situ, as it may be nutrient loss and later decrease in soil productivity, but it also may cause downstream effects as for example contributing with the process of eutrophication of watercourses or carrying agrochemicals compounds (such as pesticides or fertilizers) into surface runoff. This last issue might carry out several consequences in the watershed of El Pescado Creek, where crops and greenhouses are much more concentrated in the upper watershed meanwhile recreational activities associated to the watercourse are often developed in the middle and lower watershed (fishing, kayak). Therefore, we conclude that this study, although being a preliminary approach, might be considered as a useful tool for the management of socio-economic growth in view of the minimal territorial planning carried out in the watershed. Comparing soil loss between different land-use and land-cover scenarios enabled us to identify priority areas where there is a risk of soil erosion, in the current scenario. Further studies should include field campaigns and soil loss plots to track changes in soil erosion rates over time, bringing the chance to implement management measures before reaching a critical situation.