Fig 3 - uploaded by Beata Houskova
Content may be subject to copyright.
Comparison between measured bulk density and bulk density values calculated by pedotransfer models (BD PTF 1 , BD PTF 2 , BD B , BD MJ ).
Source publication
Soil bulk density is one of the main direct indicators of soil health, and is an important aspect of models for determining agroecosystem services potential. By way of applying multi-regression methods, we have created a distributed prediction of soil bulk density used subsequently for topsoil carbon stock estimation. The soil data used for this st...
Contexts in source publication
Context 1
... Stagnosols, Planosols, Fluvisols, Chernozems, Luvisols) with a rather wide range of agro- chemical properties. A comparison of measured bulk density (BD) and soil bulk density as calculated by the pedotransfer equations (BD PTF 1 and BD PTF 2 ) following the updated input parameters that have been evaluated in the PMS-S database (Table 5, Fig. 3) were made. Bulk density values calculated according to Bernoux et al. (1998) -BD B and to Manrique and Jones (1991) BD MJ (recommended by JRC for Europe) for ...
Context 2
... sites 6 to 8 are utilized as permanent grass- lands, the rest of the sampling sites are of arable land. We found that models based on pedotransfer functions generally slightly lower the value of bulk density (with the excep- tion of sampling sites 3, 9, 10 and 13 (Fig. 3)). Moreover, the best prediction of BD in the set of key sampling sites was from the BD PTF 1 model, according to the average values, together with minimum and maximum values ( Table 6). The differences between measured soil bulk den- sity and the model values vary from -0.144 to 0.243 g cm -3 . Furthermore, the average value of ...
Similar publications
Land use and management practices affect the response of soil organic carbon (C) to global change. Process-based models of soil C are useful tools to simulate C dynamics, but it is important to bridge any disconnect that exists between the data used to inform the models and the processes that they depict. To minimise that disconnect, we developed a...
The shrub Nitraria tangutorum is distributed widely in arid desert areas, and plays a critical role in the desert–oasis ecosystem. This study quantified varying characteristics of carbon (C) and nitrogen (N) in the soil at four stages—the initial stage (IS), stable stage (SS), degradation stage (DS), and severe degradation stage (SDS)—in a steppe e...
The conversion of lowland tropical forests to oil palm plantations is known to reduce the amount of soil carbon and nitrogen. However, this has been criticised as the plantations also contribute to a net gain of the soil carbon (C) stock. In this study, carbon and nitrogen stable isotopes techniques was used to confirm the source of carbon stored i...
Incorporating native perennial grasses adjacent to annual row crop systems managed on marginal lands can increase system resiliency by diversifying food and energy production. This study evaluated (1) soil organic C (SOC) and total N stocks (TN) under warm-season grass (WSG) monocultures and a low diversity mixture compared to an adjacent no-till c...
Plain Language Summary
Soil contains a lot of carbon, and its levels can change based on climate and how we manage the land. Our global soil research shows that when new carbon goes into the soil, about 25% of it leaves within a year, regardless of depth. After 30 years, only about 13% remains. Interestingly, most of the carbon in the soil is older...
Citations
... These functions can be obtained from mathematical models based on easily obtained edaphic attributes (Andrade et al., 2020), allowing streamlining and optimization of the evaluation of attributes that are difficult to obtain, such as soil bulk density (Bd). Bd can present great spatial variability, and its quantification presents a challenge in soil science (Makovníková et al., 2017). Bd is related to soil porosity, which directly influences gas exchange, resistance to root system development, water circulation in the soil, and consequently, the intensity of the erosive process. ...
Objective: This study aimed to obtain the Bd of organic constitution horizons through PTFs and evaluate the C and N stocks of different soil classes found in the upper part of the INP. Theoretical framework: The evaluation of bulk density (Bd) depends on the method used to obtain the sample and presents great spatial variability, especially in mountainous regions such as the upper part of Itatiaia National Park (INP), Brazil. Techniques can be used for this purpose, including those of pedotransfer functions (PTFs) allowing the quantification of nutrients such as C and N in soils. Methodology: The research was conducted within the Itatiaia National Park (INP), situated in the Mantiqueira Mountains of southeastern Brazil. Soil samples were collected from the upper portion of the INP, specifically targeting soil profiles containing hystic horizons (O or H). The study employed pedotransfer functions (PTFs) established by Beutler et al. (2017) to estimate the soil bulk density (Bd) of organic horizons. Based on these assessments, stocks of carbon (C) and nitrogen (N) were calculated. The standard depth for these calculations was 30 cm. In total, 47 profiles were selected, 26 of which belonged to the order Histosols, 15 to Cambisols, and 6 Leptosols & Regosols. Results and conclusion: In general, a high C content, between 5 and 15 g kg-1, and N contents greater than 1.0 g kg-1 were observed. For Bd an increase in depth occurred, with values close to 0.50 Mg m-3 in the superficial horizons and > 0.75 Mg m-3 in depths greater than 30 cm. The area occupied by soil profiles presenting hystic horizons corresponds to 16403.69 ha (hectares) with 315645.7 Mg of C and 20043.0 Mg of N. Of this total, 10254.44 ha were Cambisols that store approximately 205282.1 Mg of C and 13121.8 Mg of N. Histosols occupied an area of 4148.38 ha and stored 90698.4 Mg of C and 5576.0 Mg of N. The area occupied by Leptosols & Regosols was 1347.74 ha and they stored 19665.1 Mg of C and 1345.1 Mg of N. For the Cambisols, the average values of C and N stocks were lower than those quantified in the Histosols, but considering the area occupied, the total C stock was more than twice the amount quantified in the Histosols. Research implications: Soils with organic horizons complicate Bd determination due to costly and unrepresentative collection methods. Additionally, converting soil masses to volume for nutrient stock measurements (e.g., C and N) is problematic. Due to limited accessibility in certain areas, obtaining undisturbed soil samples for evaluation can be challenging. In such cases, pedotransfer functions (PTFs) play a crucial role in assessing and modeling soil phenomena. Indirect quantification of Bd using PTFs has become essential for addressing these issues. Originality/value: This study utilizes pedotransfer functions (PTFs) to estimate soil bulk density (Bd) and, consequently, nutrient stocks in fragile mountainous soils, enabling monitoring of these environments.
... Soil is a natural resource consisting of minerals and organic matter that are bound by mineral oxides and charged clay particles. There are spaces between particles that allow the exchange of air and water [1]. The number and size of these pores vary indicating variation in organic matter content, quality, and formation. ...
This study was undertaken to determine the effect of Soil pH, porosity, bulk and
particle density, magnesium, manganese, sodium, copper, zinc, and titanium on
the physicochemical and nutritional content of Syzygium cumini fruits.
Multistage sampling was used to select the counties, identify, sample and
collect fruit trees. Analysis was done using 40 soil and fruit samples each per
county using standard methodologies. Fruit parameters were evaluated by
measuring juice pH, total soluble solids, titratable acids, vitamin C, crude fat and
fiber, proteins, fruit maturity, carbohydrates, energy, sodium, magnesium,
manganese, calcium, iron, copper, and zinc. Results showed that, Syzygium
fruits preferred a soil pH from 5.3 to 6.9 which was significantly different
between the two counties. Titanium was significantly high in Kwale soil samples
giving 57.53±8.37 mg/kg. Particle density caused an increase in bulk density,
which in turn increased fruit weight, pH, and ash content. Proteins had a weak
positive correlation with soil magnesium, 0.11 and very strongly correlated with
carbohydrates, 0.99. Mature fruits had significantly higher protein and
carbohydrate contents of 2.96 and 36.1 mg/100 g respectively. The
physicochemical and nutritional content of Syzygium cumini fruits were highly
influenced by soil properties.
... There are several methods for determining compaction in the soil, including the measurement of the soil bulk density (Upadhyaya et al. 1982;Raper et al. 1990). Soil bulk density is one of the main and direct indicators of soil health (Abbott and Manning 2015), and is one of the most important parameters for assessing the soil sensitivity to compaction and compaction intensity (Makovníková et al. 2017). In addition, soil bulk density is a dynamic property as it changes with time and space. ...
... Their results showed that the extended PTFs using artificial neural network (ANN-PTFs) outperformed the extended PTFs by multiple linear regression (MLR-PTFs). Makovníková et al. (2017) used the multi-regression methods to predict the soil bulk density using soil data acquired from the Slovak soil database by considering the SOC content and soil size distribution as independent input parameters. Their results showed that the difference between the measured values and the predicted values varied from -0.144 g/cm3 to 0.135 g/cm3. ...
... The neural network models developed in this study showed the absolute superiority of the ANN model for predicting the soil bulk density over regression models (e.g., Ruehlmann and Körschens 2009). The advantage of ANN is indicated by comparing the statistical parameters of the neural network model (R 2 = 0.93, MSE = 0.038) with those of the regression model proposed by Ruehlmann and Körschens (2009) (R 2 = 0.21, RMSE = 0.142) and Makovníková et al. (2017) (R 2 = 0.46). A comparison of the neural system created in our study with the ANN models generated by Xiangsheng et al. (2016) (R 2 = 0.71, RMSE = 0.14) and Al-Qinna and Jaber (2013) (R 2 = 0.61, RMSE = 0.106) also confirms the superiority of our neural model applying the soil cone index (CI), soil moisture content (MC), and soil electrical conductivity (EC). ...
Developing models that can accurately predict the soil bulk density using other soil parameters is of great importance given the arduous task of determining the soil bulk density. We conducted the factorial experiment based on the randomized complete block design with five replications to determine the factors that affect the soil bulk density (BD) in three types of soil texture: loam, sandy loan, and loamy sand. Artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were applied to predict and model the soil bulk density using several independent parameters that affect it, including the soil cone index (CI), moisture content (MC), and electrical conductivity (EC). The analysis indicated that the developed ANN using the Bayesian tuning algorithm with R2 = 0.93 is the most suitable model compared to other models that were created. We also performed the soil bulk density modeling using the three effective parameters of soil by ANFIS (adaptive neuro-fuzzy inference system) applying the hybrid method. The coefficient of determination for the ANFIS model was 0.988 (R2 = 0.98), which indicates the correct choice of the parameters affecting the soil bulk density. The comparison between the artificial neural network models and the neuro-fuzzy model developed in this study shows the complete advantage of ANFIS systems in predicting the soil bulk density as supported by the statistical parameters The results showed that ANN and ANFIS are highly capable to predict the soil bulk density in agricultural lands.
... Direct measurement of bulk density, depending on the country, is quite widely used in agricultural research to determine the compaction of soils, primarily of arable land, observed under different tillage systems. In general, this problem is acute both in Ukraine (Medvedev 1997a, Medvedev et al. 2004) and in other countries (Jalabert et al. 2010, Makovníková et al. 2017). The problem of agrophysical degradation of soils, along with other types of it, threatens the sustainable use of soils and the fulfillment of their global functions. ...
... Compaction led to the deterioration of other agronomically and ecologically important soil properties. The dependence of the soil bulk density on human production activity is manifested not only in how and when the soil is cultivated, but also in the content of humus (Makovníková et al. 2017, Meurer et al. 2020, Taalab 2013 or features of the indicators of the soil absorption complex (Cherlinka 2001, Nazarenko et al. 1998, 2000. In particular, the peptization of the finely dispersed part of the soil or, conversely, its aggregation depends on the composition of exchangeable cations, which is reflected in the value of the soil bulk density (Manrique andJones 1991, Heuscher et al. 2005). ...
... Since in some cases the connection can be quite transformative, its evaluation is a separate task (Kaur et al. 2002, Nazarenko et al. 1998, Sequeira 2014, Suuster et al. 2011, Tranter et al. 2007. For example, Nazarenko et al. (1998) found that for different soil types the relationship between soil bulk density and humus content is inversely proportional and is both linear and non-linear which is often confirmed by other authors (Chestnykh and Zamolodchikov 2004, Demakov et al. 2017, Makovníková et al. 2017, Manrique and Jones 1991. The obvious influence of the humus content is not only on the bulk density but also on soil aggregation, and physical and hydraulic properties (Makovníková et al. 2017, Meurer et al. 2020. ...
Soil monitoring programs in many countries often do not measure such an indicator as soil bulk density. At the same time, modern problems of assessing the condition of the soil cover require its use in soil management of agricultural landscapes. We see the solution to this problem in geospatial modeling. We have developed a technique for simulating soil bulk density, the algorithm of which consists of five stages: the creation of model maps of agro-industrial soil groups, modeling of a clay content map, construction of a map of humus content, selection of empirical dependences of soil bulk density on humus and clay content with evaluation reliability of these dependencies and their implementation in GRASS GIS and, finally, cartographic modeling of the spatial continuous distribution of soil bulk density. At the same time, parameters for improving the quality of methods modeling and mapping the soil bulk density were established. First of all, this is an increase in the sampling dataset of density, humus, and clay content to improve the reliability of the selected regression formula. It is also an improvement in the quality of spatial modeling of the humus content using geostatistical methods or modeling. To improve the clay content map, it is necessary to increase the resolution of DEM and predictors, to choose more accurate sources of DEM and prediction algorithms. The application of the presented methodology is shown in a part of Chernivtsi region, Ukraine, with the possibility of use on the scale of the entire country. URL: https://www.gcass.science.upjs.sk/
... In an ideal soil, solid components provide root growth medium, attachment and nutrients for plants, while pore spaces provide the air and water needed (Aşkın & Özdemir, 2003). BD which is one of the important indicators of soil quality (Abbott & Manning, 2015) is closely related to environmental quality and biomass production (Makovníková et al., 2017). BD is a dynamic soil property, as it varies in space and time. ...
... BD is a dynamic soil property, as it varies in space and time. It is effected by land and crop management practices (Çerçioğlu et al., 2019; Çerçioğlu, 2020), as well as by natural processes such as the climate conditions that influence soil cover, SOM contents, porosity or soil structure (Makovníková et al., 2017). Changes in BD depending on the effectiveness of the degrading and forming processes in the soil are closely related to SOM content (Demir et al., 2019;Demir & Işık, 2019, 2020Demir, 2020) and textural structure (Aşkın & Özdemir, 2003;Makovníková et al., 2017). ...
... It is effected by land and crop management practices (Çerçioğlu et al., 2019; Çerçioğlu, 2020), as well as by natural processes such as the climate conditions that influence soil cover, SOM contents, porosity or soil structure (Makovníková et al., 2017). Changes in BD depending on the effectiveness of the degrading and forming processes in the soil are closely related to SOM content (Demir et al., 2019;Demir & Işık, 2019, 2020Demir, 2020) and textural structure (Aşkın & Özdemir, 2003;Makovníková et al., 2017). BD increases with soil depth since subsurface layers are more compacted and have less aggregation, less SOM and less root penetration compared to surface layers, therefore contain less pore space. ...
In this study, the changes of some soil physical and chemical properties were investigated under different land use conditions in Turhal, Turkey. Soil samples were collected from 0-20 cm depth from twenty four plots under eight different land uses which are sunflower, wheat, vegetable, orchard, sugar beet, meadow, pasture and alfalfa plants. Some soil properties where these plants are grown and their effects on the bulk density were investigated. The findings show that basic soil properties and practices related to plant management are effective on the bulk density. While the lowest mean bulk density value was determined in meadow (1.00 g cm-3) areas, the highest bulk density value was determined in soils cultivated with sugar beet (1.71 g cm-3). Correlations between the investigated parameters were tested with the use of Pearson's correlation method. Bulk density and some soil parameters used in the evaluation of structural stability and sensitivity to erosion were found significant relationships.
... PSD impacts several soil processes, including water movement and soil degradation, and consequently influences soil quality and productivity (Curcio et al., 2013). In addition, soil texture and PSD are commonly used as inputs in water flow and contaminant transport models; they are also considered when predicting soil hydraulic properties using pedotransfer functions (Contreras and Bonilla, 2018;Makovníková et al., 2017). Most PSD analysis methods, such as the hydrometer method (Bouyoucos, 1927) and pipette method (Robinson, 1922), are based on report specific point measurements. ...
... First, the amount of affected topsoil per ha was calculated based on the LUCAS (https://esdac.jrc.ec.europa.eu/projects/lucas (24 October 2021)) soil bulk density data from Ballabio [75] and Makovinikova [76]. Bulk density of soil in EU agricultural areas varies from 1.2 to 1.4 tons/m 3 ; subsequently, a value of 1.3 was used here. ...
The pig and poultry industries continue to grow across the world and together they provide the majority of meat consumed. The European Union (EU) in particular has the highest global relative meat production by monogastrics (i.e., pig and poultry). The fate of phosphorus (P) in pig and poultry farming was studied, accounting for P content in feed, animals, manure, soil, and runoff. P input from manure, and P offtake in crops receiving manure, were plotted against each other to arrive at “safe” P loading rates, in order to minimize soil P surpluses along the lines of the EU Nitrogen Expert Panel in their work with nitrogen (N). However, it was observed that it is the N/P ratio and the background soil P levels that determine whether a certain manure will end up producing surplus levels of soil P. Critical N/P weight ratios were derived over different crop P offtake rates when applying stored manure to croplands. At spreading rates of 170 and 250 kgN/ha/year and a crop P offtake of 15 or 30 kgP/ha/year, stored pig and chicken manure result in soil P surpluses. An important factor in determining effective N/P ratios is the plant availability of N in stored manure, which runs at around 47%, estimated from previously published results. The minimization of N losses to the atmosphere and to groundwater in housing, storage, and spreading of manure has a major impact on the N/P weight ratio of the manure that ends up on fields. In most cases, half of the ex-animal N content has been lost in stored or degraded manure, with N/P weight ratios running at two and less. Following only the EU Nitrates Directive, which allows for a maximum of 170 kgN/ha/year in NVZs (Nitrate Vulnerable Zones), will often result in soil P surpluses leading to runoff losses to adjacent water bodies. Therefore, for the pig and poultry industries to continue thriving, measures are required to better manage manure, including improved storage and spreading techniques, acidification, separation, struvite extraction and ammonia stripping of pig slurry, and drying and pelleting of poultry litter. This way, excess manure and derived biofertilizers from animal farms can find their way back into the commercial market, instead of ending up as legacy P in watersheds and coastal zones.
... Since then, other PTFs for BD estimation have been developed based on the fine earth fractions and SOC, which is important to BD due to its effect on the ratio between soil macro-and micropores (Martín et al., 2017;Throop et al., 2012). Furthermore, many other functions have been proposed to describe regional (Akpa et al., 2016;Chagas et al., 2016;Chen et al., 2018;Makovníková et al., 2017;Montzka et al., 2017;Ramcharan et al., 2017;Román Dobarco et al., 2019;Wösten et al., 1999Wösten et al., , 2013 and local conditions (Benites et al., 2007;De Vos et al., 2005;Picciafuoco et al., 2019;Sevastas et al., 2018) and to predict BD in different soil horizons (Hollis et al., 2012;Reidy et al., 2016;Sequeira et al., 2014). ...
For the estimation of the soil organic carbon stocks, bulk density (BD) is a fundamental parameter but measured data are usually not available especially when dealing with legacy soil data. It is possible to estimate BD by applying pedotransfer function (PTF). We applied different estimation methods with the aim to define a suitable PTF for BD of arable land for the Mediterranean Basin, which has peculiar climate features that may influence the soil carbon sequestration. To improve the existing BD estimation methods, we used a set of public climatic and topographic data along with the soil texture and organic carbon data. The present work consisted of the following steps: i) development of three PTFs models separately for top (0–0.4 m) and subsoil (0.4–1.2 m), ii) a 10-fold cross-validation, iii) model transferability using an external dataset derived from published data. The development of the new PTFs was based on the training dataset consisting of World Soil Information Service (WoSIS) soil profile data, climatic data from WorldClim at 1 km spatial resolution and Shuttle Radar Topography Mission (SRTM) digital elevation model at 30 m spatial resolution. The three PTFs models were developed using: Multiple Linear Regression stepwise (MLR-S), Multiple Linear Regression backward stepwise (MLR-BS), and Artificial Neural Network (ANN). The predictions of the newly developed PTFs were compared with the BD calculated using the PTF proposed by Manrique and Jones (MJ) and the modelled BD derived from the global SoilGrids dataset. For the topsoil training dataset (N = 129), MLR-S, MLR-BS and ANN had a R2 0.35, 0.58 and 0.86, respectively. For the model transferability, the three PTFs applied to the external topsoil dataset (N = 59), achieved R2 values of 0.06, 0.03 and 0.41. For the subsoil training dataset (N = 180), MLR-S, MLR-BS and ANN the R2 values were 0.36, 0.46 and 0.83, respectively. When applied to the external subsoil dataset (N = 29), the R2 values were 0.05, 0.06 and 0.41. The cross-validation for both top and subsoil dataset, resulted in an intermediate performance compared to calibration and validation with the external dataset. The new ANN PTF outperformed MLR-S, MLR-BS, MJ and SoilGrids approaches for estimating BD. Further improvements may be achieved by additionally considering the time of sampling, agricultural soil management and cultivation practices in predictive models.
... The bulk density is an indicator of the soil compaction and health and its values depend on the texture, organic matter content, constituent minerals and porosity (Baver et al. 1972;Hanks & Ashcroft 1980;Šimečková et al. 2016). It has an effect on the root development and crop yield and usually increases with the soil depth (Dam et al. 2005;Makovníková et al. 2017). The bulk density is an important conversion of weight-based data to volume (and area) -related data (Brady & Weil 2002). ...
Soil properties can be influenced by long-term agricultural management practices as described in pedological literature. In this study, selected physical properties (particle density and bulk density, total porosity, maximum capillary water capacity, minimum air capacity, field capacity, permanent wilting point and available water capacity) of topsoils from different reference soil groups (Cambisols, Luvisols, Fluvisols, Chernozems and Phaeozems, Leptosols, Stagnosols and Gleysols) were sampled and analysed in the years 2016–2017. The topsoil samples were taken from points of so-called S (specific) soil pits to be sampled from the General Soil Survey of Agricultural Soils (GSSAS) which was accomplished in the years 1961–1970. In addition, some of the properties were also compared with those measured during the GSSAS. Recognising the properties, only the particle density, the maximum capillary water capacity, the permanent wilting point and the available water capacity of the topsoil of the individual soil groups were statistically significantly (P < 0.05) different. A comparison of the physical properties with those analysed after more than 40 years was performed, the bulk density increased and the total porosity decreased in the topsoil of the major part of the studied soil groups.
... Since then, other PTFs for BD estimation have been developed based on the fine earth fractions and SOC, which is important to BD due to its effect on the ratio between soil macro-and micropores (Martín et al., 2017;Throop et al., 2012). Furthermore, many other functions have been proposed to describe regional (Akpa et al., 2016;Chagas et al., 2016;Chen et al., 2018;Makovníková et al., 2017;Montzka et al., 2017;Ramcharan et al., 2017;Román Dobarco et al., 2019;Wösten et al., 1999Wösten et al., , 2013 and local conditions (Benites et al., 2007;De Vos et al., 2005;Picciafuoco et al., 2019;Sevastas et al., 2018) and to predict BD in different soil horizons (Hollis et al., 2012;Reidy et al., 2016;Sequeira et al., 2014). ...
For the estimation of the soil organic carbon stocks, bulk density (BD) is a fundamental parameter but measured data are usually not available especially when dealing with legacy soil data. It is possible to estimate BD by applying pedotransfer function (PTF). We applied different estimation methods with the aim to define a suitable PTF for BD of arable land for the Mediterranean Basin, which has peculiar climate features that may influence the soil carbon sequestration. To improve the existing BD estimation methods, we used a set of public climatic and topographic data along with the soil texture and organic carbon data. The present work consisted of the following steps: i) development of three PTFs models separately for top (0–0.4 m) and subsoil (0.4–1.2 m), ii) a 10-fold cross-validation, iii) model transferability using an external dataset derived from published data.
The development of the new PTFs was based on the training dataset consisting of World Soil Information Service (WoSIS) soil profile data, climatic data from WorldClim at 1 km spatial resolution and Shuttle Radar Topography Mission (SRTM) digital elevation model at 30 m spatial resolution.
The three PTFs models were developed using: Multiple Linear Regression stepwise (MLR-S), Multiple Linear Regression backward stepwise (MLR-BS), and Artificial Neural Network (ANN).
The predictions of the newly developed PTFs were compared with the BD calculated using the PTF proposed by Manrique and Jones (MJ) and the modelled BD derived from the global SoilGrids dataset.
For the topsoil training dataset (N = 129), MLR-S, MLR-BS and ANN had a R² 0.35, 0.58 and 0.86, respectively. For the model transferability, the three PTFs applied to the external topsoil dataset (N = 59), achieved R² values of 0.06, 0.03 and 0.41. For the subsoil training dataset (N = 180), MLR-S, MLR-BS and ANN the R² values were 0.36, 0.46 and 0.83, respectively. When applied to the external subsoil dataset (N = 29), the R² values were 0.05, 0.06 and 0.41. The cross-validation for both top and subsoil dataset, resulted in an intermediate performance compared to calibration and validation with the external dataset. The new ANN PTF outperformed MLR-S, MLR-BS, MJ and SoilGrids approaches for estimating BD. Further improvements may be achieved by additionally considering the time of sampling, agricultural soil management and cultivation practices in predictive models.
