SOIL EROSION SUSCEPTIBILITY EVALUATION BASED ON GIS TECHNOLOGY 
Pinto, S.A.F.(1); Valerio Filho, M.(1) 
P.L. (2) 
(1) Instituto Nacional de Pesquisas Espaciais 
12.201 - Sáo José dos Campos, C.P. 515 
Säo Paulo, 
Donzeli, 
Brasil 
(2) Instituto Agronômico de Campinas 
13.100 - Campinas, S.P., C.P. 28 
Brasil 
ABSTRACT: 
This study was undertaken in an agricultural watershed 
and its purpose was to indicate soil erosion susceptibility 
State, Brasil) 
(eastern Sao Paulo 
areas and to provide information for rural planning. A geographic information 
system - GIS (SGI/INPE) was used to integrate physical parameters of the USLE 
model adjusted. Erosivity, 
parameters (R, K, 
soil loss 
erodibility, 
L and S USLE factors, 
tolerance data to define tolerable cover-management and soil 
slope length and slope gradient 
respectively) were associated with 
conservation practices (tolerable CP factors - USLE). Tolerable CP and actual 
CP data (determined from Landsat and field informations) were integrated in 
the GIS environment. A susceptibility soil erosion classe map 
(at 1:60.000 
scale) was obtained and it will be used for soil conservation planning of the 
watershed. 
KEY WORDS: Remote Sensing, GIS, Soil Erosion 
1. INTRODUCTION 
Development countries have an urgent need 
to improve their agricultural production. 
This process induces a lot of inadequate 
land use/cover for the rural environment. 
The intensification of agricultural 
activities may result in increasing 
erosion processes and accelerated soil 
losses, threatening natural resources 
integrity, mainly water quality, and the 
productivity of agricultural systems. In 
this context, the knowledge of potencial 
erosion of specific soil groups is very 
important to support agricultural and 
environmental planning. 
GIS constitutes a technique designed to 
acquisition, storage, manipulation and 
analysis of large amount of geocoded data 
(Marble and  Peuquet, 1983; Bocco and 
Valenzuela, 1 988; Ventura et al., 1988). 
The objetive of this study is to evaluate 
the soil erosion susceptibility by using 
GIS technology in analytic integration of 
environmental data. The analysis was 
orientated by the Universal Soil Loss 
Equation model  USLE (Wischmeier and 
Smith, 1978), depicting soil loss by 
rainfall erosion. 
The selected study area is a small 
watershed at the eastern portion of Sao 
Paulo State - Brasil (Sao Joaquim river - 
229 00’- 229 05’ south lat. and 47° 20’ - 
479 35’ long west Green.), included in 
the National Watershed Management 
Program. 
2. METHODOLOGICAL PROCEDURES 
The analytical procedure was to digitally 
integrate the following USLE parameters: 
topographic factor (slope stepness and 
slope length - 18), erosivity (R), 
erodibility (K), land use and management 
(C) and conservationist practices (P). An 
172 
adjusted USLE model was applied, 
considering non-availability of different 
environmental data in Brasil (Bertoni and 
Lombardi Neto, 1985). 
Data were sampled from topographic maps 
(topographic factor), available tables 
(erosivity and erodibility) and themátic 
classification of TM and HRV imageries 
associated with field work (land 
use/management and conservationist 
pratices). 
The Natural Erosion Potential (NEP) was 
derived from the ajusted USLE as folllow: 
NEP - R * K * (0,00984 * LO:93 # 51,18) 
The information concerning NEP was 
associated with the tolerable soil loss 
level (At) to assess and spatially detect 
the  fator CP  tolerable (CPt). This 
procedure is performed due to the 
following relation spatially: 
CPt = At/NEP 
data (Se) was 
Susceptibility erosion 
evaluation of CP 
obtained through the 
actual and CP tolerable 
Se = CP actual - CP tolerable 
Data analysis for the evaluation of NEP 
and characterization of soil loss 
susceptibility were made using raster 
format data, with the support of a GIS 
developed at INPE (SGI/INPE, Souza et 
al., 1990). An erosion susceptibility map 
at scale of 1:60.000 was obtained using a 
graphic plotter. 
3. RESULTS 
The observed C factor was derived from a 
land use map obtained through TM and 
HRV/SPOT imageries classification (Figure 
1). The P factor was considered as a 
constant value for the whole watershed, 
  
  
  
rep 
0,5 
wer 
val 
the 
ove: 
coi 
slo; 
Sus: 
the 
cor: 
tol: 
dat: 
nul: 
sus 
sho 
the 
The 
Sus 
res] 
are 
ero: 
Neve 
NEP 
soi: 
pro 
vege