SOIL EROSION MODELLING WORKSHOP

JRC – Joint Research Centre20 – 22 March 2017 – Ispra, Italy

Database for Soil

Erosion Modelling in

Brazil

Department of Soil Science

Federal University of Lavras

National Counsil of Technological and Scientific Development

Foundation for Research Support of Minas Gerais

Prof. D.Sc. Marx Leandro Naves SilvaSoil and Water Conservation

marx@dcs.ufla.br

Research Fellow of CNPqSoil Physic and Soil and Water Conservation Lab

1Introduction

Brazil is one of the most important food, fiber and fuel producer in the world.

Expansion of cropland and pasture areas may lead to increased soil erosion rates.

Brazilian scientific research on soil erosion has a relatively recent formation,

although the first works were published in the forties, and approximately half of the

scientific production originated in the last fifteen years.

Observed annual soil loss in soil erosion plots ranges from 0.1 to 175.4 ton ha yr−1.

In Brazil, it is estimated that there are more than one hundred plot experiments for

studies of soil erosion.

Efforts should continue to allow field observations in all regions of Brazil, mainly

where data is scarce (central-western and northern regions).

Models for monitoring and estimating soil erosion used in Brazil are: USLE (Universal

Soil Loss Equation), RUSLE (Revised Universal Soil Loss Equation), WEPP (Water Erosion

Prediction Project), SWAT (Soil and Water Assesment Tool).

Denardin (1990); Silva et al. (1997); Marques et al. (1997); Oliveira et al. (2008); Silva et al. (2009) Anache et al. (2017)

2What is the best model for Brazil?

We studied soil loss and soil loss tolerance spatialization,estimated by RUSLE versus WEPP models in RS – Brazil.

Silva et al. (2014)

The models presented different values of soil losses.

3Rainfall Erosivity – R FACTOR

In Brazil, there are 73 regression

equations to calculate erosivity.

These studies are concentrated

in the South and Southeast regions

(~60%) with a few studies in the

other regions, mainly in the North.

The annual rainfall erosivity in

Brazil ranges from 1,672 to 22,452 MJ

mm ha−1 h−1 yr−1.

Techniques already established

in Brazil may be used for the

interpolation of rainfall erosivity,

such as geostatistics and artificial

neural networks.

Oliveira et al. (2012)

4Rainfall erosivity in small and large scales

The watershed of the Jaguarí and

Camanducaia River, MG

(Pontes et al., 2015).

The watershed of the Alto Rio

Grande River, MG

(Carvalho et al., 2015).

The watershed of the Doce River, MG (Silva et al., 2010).

Rainfall erosivity map of Brazil, an

approximation (Oliveira et al., 2012).These watersheds are of importance for the production of drinking

water, electric power production and mining tailings retention dam.

5Soil Erodibility – K FACTOR

DATABASE: soil erodibility obtained directly: plots under natural and

simulated rainfall = 61 soils. The erodibility values show an amplitude of

0.0004 to 0.0840, with the mean value of 0.0185 Mg ha h ha-1 MJ-1 mm-1.

Mondardo et al. (1978); Resck et al. (1981); Dal Conte (1982); Angulo (1983); Távora et al. (1985); Bertoni & Lombardi Neto(1985); Rodrigues do Ó (1986); Silva et al. (1986); Fernandez Medina & Oliveira Júnior (1987); Carvalho et al. (1989);Denardin (1990); Levien, citado por Denardin (1990); Leprun, citado por Denardin (1990); Campos Filho et al. (1992);Martins Filho & Pereira (1993); Silva & Andrade (1994); Silva (1994); Silva et al. (1994); Carvalho et al. (1994); EMBRAPA-CNPF(informação pessoal do pesquisador G.R. Curcio); Marques (1996); Silva (1997); Hernani et al. (1997); Marques et al. (1997);Silva et al. (1997); Bertol et al. (2002); Martins et al. (2011); Eduardo et al. (2013); Corrêa (2016); Silva et al. (2016).

All soil classes were consideredID Soil K-Factor

1 RL 0,005

2 PVA 0,032

3 PVA 0,031

4 PA 0,007

5 PA 0,0004

6 FX 0,017

7 LVA 0,002

8 LV 0,026

9 SN 0,012

10 LV 0,012

11 NL 0,008

12 TC 0,009

13 LV 0,016

14 LV 0,004

15 PV 0,034

16 PV 0,0026

17 RR 0,002

18 PA 0,045

19 PVA 0,014

20 LV 0,009

21 ESK 0,032

22 TC 0,032

23 LV 0,009

24 CX 0,084

26 PV 0,028

27 PV 0,018

28 LV 0,009

29 NC 0,044

30 CH 0,011

31 LV 0,004

32 LVA 0,010

33 CX 0,035

34 LV 0,025

35 LA 0,011

36 PVA 0,023

37 PVA 0,004

38 LV 0,008

39 LV 0,021

40 TC 0,008

41 PV 0,004

42 RL 0,008

43 PV 0,004

44 LV 0,013

45 LV 0,022

46 VX 0,006

47 PVA 0,008

48 CX 0,047

49 PVA 0,061

50 PV 0,032

51 PV 0,024

52 TR 0,012

53 PVA 0,009

54 PVA 0,033

55 LV 0,002

56 PVA 0,025

57 PV 0,008

58 RQ 0,003

59 LA 0,009

60 LVA 0,034

61 PVA 0,027

1:5.000.000

6Soil Erodibility – Ind i rect Methods

MODELS SUGGESTIONS:

Denardin, (D. Sc. Thesis): general soil classes.

Marques et al. (RBCS, 21:447-456, 1997): Soils with

Textural Horizon B (Argisols, Nitosols, Planosols and

Luvisols)

Silva et al. (PAB, 35(6):1207-1220, 2000): (Latosols)

Oxidic Latosols (Oxisols) have high clay content,

however, these soils have low erodibility due to their

granular structure, which increases water infiltration.

The structure is related to the soil’s oxidic mineralogy.

Hence, indirect methods must be adapted.

7Topographic Factor – LS FACTOR

Development of topographic factor modelling for application in soil erosion models:

USPED, RUSLE 3D (Mitasova et al. 1996), and Desmet & Govers (1996)

The complexity of topography is considered

It determines the upstream area that contributes to the flow along each point in the

basin, considering the effects of flow convergence/divergence

Large-scale landscape analyses

Reduction of subjectivity: systematic analyses based on Geographic Information

Systems (GIS) tools

Interpretation of the researcher: adaptations of analog techniques for the empirical

processing of the data

Ease and speed in data processing

Lower relative cost

Oliveira et al (2013)

8RUSLE versus RUSLE 3D – LS FACTOR

Maps of the LS factor obtained by RUSLE and RUSLE 3D methods in RS – Brazil.

Silva et al (2014)

Different values for the LS factor generate doubts.

9Cover and Management – C FACTOR

10Conservation Practices – P FACTOR

Variables involved in obtaining cover factor: vegetation development;

crop management; crop spacing; root system; plant architecture; plant

physiology; deposition and decomposition of residues.

The soil cover factor is among the most difficult factors to be obtained

by field or laboratory experimentation.

Another important factor in the modelling of soil erosion is the

conservationist practices related to the control of soil erosion.

In Brazil, studies related to this factor are still rare.

DIFFICULTIES: cost; time-consuming in the experimental conduction;

long-term studies; large-scale studies.

11Land Use in Brazil (IBGE, 2014) and C Factor

Land UseContribution

%Millions ofhectares

Native forests and pastures 66.1 562.5

Planted forests 0.9 7.7

Planted pastures 23.3 198.3

Agriculture 6.2 52.8

Urban 3.5 29.8

Total 100 851.0

VEGETAL COVER C - FACTOR VALUES SOURCE

Soybean 0.00700 – 0.21930 Bertoni et al. (1975), Bertol et al. (2001), Amaral (2006)

Wheat 0.03500 – 0.21580 Bertol et al. (2001), Amaral (2006)

Corn 0.00700 – 0.15600 De Maria & Lombardi Neto (1987), Bertol et al. (2002), Eduardo et al. (2013)

Sugarcane 0.05000 – 0.43140 Donzeli et al (1992), Cavalieri (1998), Bertoni & Lombardi Neto (2010)

Palm forage 0.25280 – 0.54280 Albuquerque et al (2007)

Cotton 0.68460 Bertoni et al. (1975)

Coffee 0.08660 – 0.65680 Rufino et al. (1985), Biscaia & Osaki (1994), Prochnow et al. (2005)

Pastures 0.00840 – 0.22000 Silva et al. (2014), Santos et al. (2016)

Eucalyptus 0.01600 – 0.30000 Martins et al. (2010), Silva et al. (2014), Silva et al. (2016)

Native Forest 0.00140 – 0.01780 Albuquerque et al. (2007), Martins et al. (2011), Santos et al. (2014), Silva et al. (2014), Silva et al. (2016)

In Brazil there is in the literature the value of the C factor for different crops and

management systems.

1:5.000.000

12Opportunities and Future Challenges

▪ Increase of the network of standard plot studies for soil and

water loss studies, notably in the North and Central-Western

regions

▪ Increase the network of automated rainfall and

sedimentological stations

▪ Development of algorithms and interpolators in the concept

of artificial intelligence

▪ Development of connectivity studies of soil erosion and river

sedimentation

▪ Use of remote sensing resources for data generation

▪ Development of model validation methodologies

THANK YOU FOR YOUR

ATTENTION!

JointResearchCentre