La vulnerabilidad frente a los desastres naturales

For decades, a basic principle guiding SOM management has been that equilibrium levels of C and N are controlled by net inputs rather than by qualitative properties (Larson et al., 1972; Rasmussen et al., 1980; Havlin et al., 1990; Paustian et al., 1992; Gregorich, 1995) and while simple differences in BNPP are certainly important, recent evidence suggests that qualitative differences in plant species are more significant than indicated by these earlier studies (Drinkwater et al., 1998; Gale & Cambardella, in press; Puget and Drinkwater, submitted). Plant species can impact SOM equilibrium through differences in the tendency to foster the formation of soil aggregates, biochemical composition of litter and through interactions with native SOM. Although our understanding of these qualitative aspects of plant-soil interactions is not fully developed, there is enough information to expect that choosing plant species

with specific traits, for example, a species that enhances aggregate stability need to be considered in designing rotations. These characteristics need to be considered along with potential BNPP.

Small grains, such as barley and rye, and legumes tend to promote aggregation to a greater extent than many cash crops (Tisdall & Oades, 1979; Reid & Goss, 1981; Angers & Mehuys, 1988, 1989). Reid & Goss (1981) found that maize, tomato and wheat actually decreased aggregate stability while growth of perennial ryegrass and lucerne tended to increase it. This increased aggregate stability has been attributed to polysaccarhides produced in the rhizosphere (Reid & Goss, 1981) and increased fungal populations associated with these species (Tisdall & Oades, 1979; Haynes & Beare, 1997). Haynes & Beare (1997) compared two legumes (white clover and lupin) and four non-legumes (Italian ryegrass, prairie grass, barley, and wheat) and found that the two legume species resulted in unexpectedly high aggregate stability relative to their root biomass suggesting that there may be fundamental differences in the rhizosphere of leguminous vs. non-leguminous plants. Increased aggregation in the legumes appeared to be related to greater fungal hyphae length in aggregates. For example, although the total root mass was similar in wheat and lupin, a greater proportion of lupin root biomass was located in aggregates and fungal hyphal length in these aggregates was 4-fold greater than in aggregates from soil where wheat was growing (Haynes & Beare, 1997).

Species differences in composition of residues have significant effects on the short-term dynamics of biodegradation and have been well-characterised in numerous microcosm experiments. Overall decomposition rate of plant residues is positively correlated with the percentage of cold-water-soluble material (sugars, free amino acids and soluble minerals), and negatively correlated with C:N ratio, lignin content, and lignin:N ratio (Taylor et al., 1989; Johansson, 1994; Prescott & Preston, 1994; Couteaux et al., 1996). The pivotal role of litter quality in controlling decomposition, N mineralisation and humus formation in natural ecosystems is generally acknowledged (Wedin & Tilman, 1990; Hobbie, 1992; Couteaux et al., 1995). These effects of individual plant species on ecosystem-level processes such as N mineralisation dynamics can be expressed in relatively short time periods in ecosystems consisting of native perennial plant species (Wedin & Tilman, 1990).

Significant differences in plant litter biochemical composition also contribute to improved retention of both C and N in agroecosystems (Drinkwater et al., 1998). Research in our long-term experiment, the Farming Systems Trial, found that residue quality differences related to plant species had a significant impact on SOM equilibrium. Over the course of 15 years, there were significant differences in organic residue inputs and in soil C sequestration in the conventional (CNV) and organic systems. Quantitative differences in residues returned and N balances across agroecosystems did not account for the observed changes in soil C and N levels (Table 1). The conventional system had greater mean cumulative aboveground net primary productivity (ANPP, Table 1) and returned a greater amount of crop residues to the soil compared to the organically managed systems. The quantity of C inputs was not the major driving force affecting soil C storage in these cropping systems. Even though the manure-based (MNR) and conventional systems received equal amounts of C, only the MNR system showed a significant increase in soil C (Table 1). The legume-based system (LEG), with lower average C inputs from aboveground sources also showed an increase in soil C. In the CNV system, aboveground inputs were in the form of senescent, high C:N crop residues, whereas both organic systems had significant inputs of low C:N residues including leguminous green manures. Changes in the 13_C

natural abundance of soil C suggest that differences in plant species composition across cropping systems have contributed to differential retention of C and N. These results indicate that the use of low C:N organic residues to maintain soil fertility, combined with greater temporal diversity in cropping sequences increased the retention of both soil C and N.

Our results differ from those reported from several classical experiments conducted in agroecosystems have concluded that residue quality related to plant species difference has no significant effect on SOM equilibrium (Larson et al., 1972; Rasmussen et al., 1980; Paustian et al., 1992; Hassink, 1995). These studies have been widely cited and have probably played a significant role in the development of C management strategies that focus on the quantity of aboveground biomass production rather than composition or source of residues. Several reported no differences in total SOM accumulation in treatments receiving straw versus leguminous residues (Rasmussen et al., 1980; Paustian et al., 1992; Hassink, 1995), however, the shoot residues were imported to these plots rather than being grown in situ eliminating important species differences in root litter quality. Given the greater impact of below ground C inputs, species differences may be expressed largely through differences in root/microbial/soil interactions.

Table 1 Cumulative C fixed by aboveground net primary productivity and C in residues returned to the soil during 1981-95 (cumulative Mg C ha-1 _{over 15 years). Net changes in soil C}

after 15 years are also given. Senescent residues are from crops and weeds. Residues incorporated as living plants were mainly legumes in the LEG system but were predominantly grasses in the MNR system. Numbers within a column followed by a different letter are significantly different at the 0.05 probability level (protected Scheffe’s). Reprinted by permission from Nature (Drinkwater et al.) copyright (1998) Macmillan Magazines Ltd.

Plant residues returned Cropping

system _{productivity Senescent Living Total} Net primary Manure input residue input Total organic C 1981 to 1995 Change in soil

MNR 69a 21 3,7 25a 19 44b 2,0

LEG 68a 31 7,5 39b 0 39a 6,6

CNV 75b _{43 0,0 43}c _{0 43}b _2,2 *

*_{There was no significant change in soil C over time in the CNV system (ANOVA, p>0.05).}

Finally, actively growing roots can also effect C storage by either increasing or decreasing the mineralisation rate of native organic matter (Helal & Sauerbeck, 1986; Liljeroth et al., 1990, 1994). The net effect of roots on mineralisation depends on plant species and soil environmental conditions such as N availability (Tate et al., 1991; Liljeroth et al., 1994). More information is needed in this area in order to identify those plant species and conditions that lead to net mineralisation of SOM.