Now we recompute the USPED model using the above described combination of map algebra and r.slope.aspect. After displaying the resulting map, we can see that the range of erosion and deposition is much lower. Some areas which had erosion still have it, but at a smaller rate, and some areas became depositional. We can compare the land use composition by creating labeled ranges for C factor maps using r.support and then running r.report for both C factor map layers.

External models can be used for more complex modeling of pollutant transport, usually involving solution of partial differential equations (Mitasova et al., 2002), see for example Path sampling modeling, coupled with GRASS. However, GRASS can be relatively easily linked to any model using libgrass9 or import and export of input data and results (Mitasova et al., 2001).

More complex, dynamic study of watershed hydrology in terms of predicting surface and subsurface flow and related phenomena can be performed using several current and older versions of hydrologic models integrated with GRASS, such as r.topmodel, r.water.fea and r.hydro.CASC2D. Use of these models requires some hydrologic background, especially familiarity with hydrologic terminology and access to input data which are not as widely available as basic GIS map layers. Use of these models is beyond the scope of this book; however, it is important to note that to fully evaluate the impact of spatial distribution of land use, this type of model is needed. They capture such important effects as the reduced velocity of water flow and higher infiltration in areas covered by dense vegetation, or increased risk of flooding due to development when vegetated area is replaced by an impervious surface, for example by a parking lot. Hydrologic modeling is also an important component of several non-point source pollution models which were linked to GRASS, such as

different crops. You can use the imported land cover map with divided fields to experiment with different crops and vegetation to find which approaches work the best.

An alternative approach would be to base the new land use plan on the erosion risk map. You can create the high erosion risk map using r.recode (net erosion/deposition D > 10, 20, 50) and assign those areas high density vegetation (C factor 0.001), while keeping the same cover everywhere else:

NOTES

1 RUSLE documents, http: www.sedlab.olemiss.edu/rusle/

2 RUSLE for ArcView,

http: abe.www.ecn.purdue.edu/~engelb/agen52 6/ gisrusle/gisrusle.html

3 Wake county GIS site

http: www.wakegov.com/county/ propertyandmapping/gisdigitaldata.htm

4 Lake Wheeler Map Section http: lnweb02.co.wake.nc.us/gis/ gismaps.nsf/0762-1619!OpenPage

5 GRASS Tutorials related Web site,

http: mpa.itc.it/grasstutor/

6 North Carolina Flood Mapping Program, http: www.ncfloodmaps.com/

7 National MUIR schema, http: www.statlab.iastate.edu/soils/ muir/schema nat.html

8 Path sampling modeling, http: skagit.meas.ncsu.edu/~helena/ publwork/GiscOO/astart.html

9 libgrass software, http: grass.itc.it/related projects.html

10 SWAT software, ftp: brcsun0.tamu.edu/pub/swat/

AGNPS and ANSWERS. Unfortunately, GRASS versions of these models are no longer supported, except for SWAT.10

Chapter 13

USING GRASS WITH OTHER OPEN SOURCE TOOLS (ft)

GRASS is one of many Free Software projects in the GIS world, however, it is the only full featured free GIS at time. A comprehensive list of more than hundred free GIS projects is available online at the FreeGIS Project Web site.1 The use, development and support of Free GIS Software is promoted at this site, as well as the use and release of publicly available geographic data. Some free GIS projects can provide additional functionality to GRASS by addressing some of its unsolved or intentional constraints.

Within this chapter, we first highlight procedures that extend the geosta-tistical analysis capabilities of GRASS. We focus on two statistics software packages, the gstat and the R project. This chapter does not try to cover the theory of geostatistics. Excellent other books on theory and applications are available, such as Cressie, 1993, Bailey and Gatrell, 1995, and Webster and Oliver, 2001. In relation to the R program it is useful to read Chambers and Hastie, 1992, Venables and Ripley, 2000, as well as Venables and Ripley, 2002.

After a brief look at GPS related software tools we close this chapter with a demonstration of fast Web mapping through UMN/MapServer linked directly to GRASS for reading GIS data from a GRASS LOCATION.

Maas river bank soil pollution data. Throughout the following sections we use the Maas river bank soil pollution data (Limburg, The Netherlands, Burrough and McDonnell, 1998). These data are provided in the gstat package and used in examples in its manual. This data set is implemented also in the R/GRASS interface package. The Maas river bank soil pollution data are sampled along the Dutch bank of the river Maas (Meuse) north of Maastricht. This is a flood plain of the river Maas, not far from where the Maas enters the Netherlands (Borgharen, Itteren, about 3 to 5 km north of Maastricht).2 The river Maas is at the north-west border of the project area, traversing the area in