5. RESULTS
Perhaps the most useful indices to result from the Input/Output section of NETWRK are the "total dependency coefficients" (Szyrmer and Ulanowicz 1987) or, more appropriately, the "indirect diets" of each taxon. By reading down the column of these dependency coefficients one notes the quantitative trophic history of material reaching that given compartment.
The total dependency coefficients (TDC’s) of the heterotrophs in the mangrove ecosystem reveal that a significant fraction of the carbon reaching many of the predators originates in the detrital compartments (Fig. 1). Macroinvertebrates such as bivalves (17) and shrimps (23) depend on the detritus compartments (92, 93 and 94) for about 95%, and on aquatic gastropods (16) and polychaetes (14) for about 90% of their sustenance. Detritus contributes a significant amount of carbon to the diets of many fishes. The dependency of mullets (52) and needlefishes (38) on detritus is more than 87%. Conversely, only very small flows link detritus with mammals that mainly feed upon herbivorous species. Such is the case for Florida panthers and bobcats (88), which get a mere 10% of their carbon via detrital pathways. Mammals such as rabbits (81) and deer (89), which feed only on primary producers, utilize no carbon whatsoever from detritus.
One may employ the total dependency matrix to test how dependent each single compartment is on the other major groups. For example, looking at how much various groups are dependent on invertebrates (compartments 12 to 32) reveals that the fishes show the highest dependencies (Fig. 2). More than 100% of the carbon reaching snooks (43), tarpon and ladyfishes (35), rays (34) and needlefishes (38) passes through invertebrates. Sharks (33) are dependent on invertebrates to the tune of 133%. (These values greater than 100% simply mean that the same carbon is spending time in more than one invertebrate compartment.)
Birds, such as cormorants (61) and pelicans (62), are dependent to a significant degree on the carbon coming from fishes (82%), but the species that depend more on fishes than any others are the mammal compartment dolphins (#90, with a dependency of 107%), and the fish compartment, barracuda (#44, 108%). (See Fig. 3).
Dependencies on the remaining three big groups (herptofauna, birds and mammals) do not exceed 20% in most cases. Among the herptofauna, the two compartments with the highest dependencies on herptofauna are owls (76) and kites and hawks (70), with dependencies of 80% and 70%, respectively (Fig. 4.) These are followed by snakes (57) with a TDC of 32%, and then minks and otters (87) and racoons (86), each showing a dependency around 15%.
Foxes receive the most carbon leaving the bird compartments (Fig. 5). Forty percent of the carbon reaching the foxes once passed through birds. The second- largest dependency coefficient on the list, eagle and osprey (71) weighs in at only 11%.
Foxes also show a high dependency upon mammal compartments (40 %), but predatory felines (88) top the list (Fig. 6). Cats depend upon other mammal species by as much as 95 %. All the other compartments depend upon mammals for less than 10% of their sustenance. (Dependency coefficients generally decline as one proceeds from invertebrates up to mammals, as expected.)
Summing the dependencies on all primary producers for each compartment, one sees that, in most instances, the result is close to 100% (Fig. 7). This indicates that the system depends mainly on internal carbon fixation and receives only a small subsidy from outside the system. The one exception is birds (#’s 60 to 79), which depend on internal carbon for less than 80% of their sustenance and import the remaining percentage of their support from outside the system.
No significant differences in indirect diets were found between wet and dry seasons.
Impact analysis is used to study both positive and negative indirect effects between compartments. In high- dimensional networks, such as this mangrove ecosystem, every pair of compartments is connected in this way, at least indirectly. Coefficients in the impact matrix represent the aggregated net (positive and negative) indirect effects between two species.
Of special interest are those compartments that function as "beneficial predators". The direct impact of a predator on its prey is obviously negative; however, the overall effect can become positive, due to the potential for compensation via multiple indirect pathways. In such cases one speaks of "beneficial predators". In the two previous networks, beneficial predation was quite abundant. In the cypress swamp, for example, the American alligator was a net benefactor to 11 of its prey items.
In the mangrove system network the number of beneficial indirect links is also quite high. During the dry season there were 208 such interactions, and the number during the wet period is slightly higher, with 218 links. The number of predators that exert a positive action on their prey are 48 and 49 for the dry and wet season respectively. Hence, at least some predators are beneficial to more than one prey type. Table 1 and table 2 list all the beneficent predators during either season, together with the number of prey that each positively impacts.
The "Snakes" compartment benefits the largest number of prey. During both dry and wet season, it exerts a net positive effect upon 14 of its prey. Other compartments that benefit large numbers of prey (nine or more) are: crocodiles (58), loons and grebes (60), pelicans (61), cormorants (62), big herons and egrets (63), small herons and egrets (64) and predatory ducks (68).
Although the number of beneficial links is high, the magnitude of the action is, in many cases, not very strong. The most significant positive impacts are given in table 3. (A significant impact is one in which more than 1% of the throughput of the prey compartment is affected. During both seasons, only nine links from among more than 200 exceed this threshold.) The strongest beneficial action (8.1%) is exerted during the wet season by the Snakes (#57) on their Lizard prey (#56). This particular interaction also imparts the greatest such benefit to the prey during the dry season, however the magnitude of the effect has dropped to 6.7%.
As mentioned above, a high number of positive actions exerted by a compartment does not necessarily indicate a strong net positive effect. For example, snakes are beneficial to 14 of their prey, but none of these effects are significant (greater than 1%). On the other hand, there is a fish compartment, sciaenids (#50), that benefits only 7 prey, but five of them are significant in the dry season and four during the wet. The compartments that benefit from the sciaenids are the aquatic gastropods (#16), bivalves (17), macrobenthos (#18), small crustaceans (#19) and amphipods (#20). (During the wet season aquatic gastropods do not benefit).
Although only a small number of beneficial predator interactions achieve significant magnitude, in a few cases more than one predator affects the same prey in a net positive manner. Adding together all the positive effects by multiple predators, one obtains an aggregate positive impact by the several predators (Table 3). In addition to those links cited above, three other prey receive significant benefit from a combination of predators during the wet season: Caridean shrimps (#22), other shrimps (#23) and predatory crabs (#28). During the dry season, the prey receiving significant combined benefits from multiple predators are four in number: polychaetes (#14), freshwater mavroinvertebrates (#29), Caridean shrimps (#22) and other shrimps (#23).
Having considered and identified instances of beneficial predation, one should also consider whether the reverse relationship might be possible, i.e., that the overall effect of a host on its predator over all pathways, direct and indirect, can be negative. In such case we talk of a "malefic prey". A search of the mangrove network uncovers 186 cases of "malefic" links during the wet season and 185 for the dry. A total of 47 prey affect their predators negatively in the wet period and 48 in the dry (Table 4). The compartment predatory crabs (#28), produces the highest number of negative indirect effects on its predators during both seasons, adversely affecting some 19 prey during the dry period.
As with most ecosystems studied by Network Analysis thus far, none of the compartments of the mangrove ecosystem feed, on the average, at or above trophic level 5. Owls feed highest on average (level 4.53 during the wet season), which is close to where the raptors in Florida Bay are feeding (level 4.59, also in the wet season.) Both of these systems stand in marked contrast to the birds in the cypress swamps, most of which feed a full trophic level lower. (Those highest in trophic rankings in the swamp were the alligators and snakes, which feed at level 3.79.) The average trophic levels during both seasons are listed in Table 5. There were almost no changes in trophic levels between seasons; the greatest difference was in the foxes, which fell from 3.93 in the wet season to 3.72 in the dry.
A quick perusal of the Lindeman transformation matrix, which apportions compartments among trophic levels, reveals that there are 19 trophic levels in the overall network. (Meaning that at least one non-redundant trophic pathway with 19 links can be found in the network.) As usual, however, not much carbon reaches beyond the fifth trophic stage, and the amounts calculated to reach the 19th level are absolutely infinetesimal (of the order 10-38 grams.) The first six trophic levels during each season are depicted on Fig. 8. One notices immediately that the trophic chains in the two seasons are virtually identical. The efficiencies of the trophic levels are just slightly lower than those observed in Florida Bay, but significantly greater (ca. 15%) than those in the cypress swamp ecosystem.
The ratio of detritivory:herbivory is a little greater than 7:1 in the mangrove. This is somewhat lower than the 10:1 ratio for temperate estuaries, but only slightly less than the 8:1 calculated for Florida Bay and higher than the 5:1 ratio in the cypress swamp.
In many areas, the mangrove ecotope lies physically between the freshwater cypress biome and the waters of Florida Bay. As both of these networks have been estimated in the previous two years of this contract, it is useful to compare the system level indices calculated for the mangrove ecosystem with those of its neighbors. It is useful to keep in mind here that in some ways the mangrove is intemediate to its neighbors, and might be expected to exhibit system indices intermediate to the other two. A word of caution is in order, however, as transition regions can be regions of greater stress than the endpoints they join. For example, estuaries, such as the mangroves, are known to subject resident populations to much higher osmotic stresses than are encountered in either fresh or seawater.
With respect to total activity, the mangrove system is almost identical to the Florida Bay community (Table 6). Its total system throughput is 3,342 gCarbon/m2/y, while that of Florida Bay is 3,459 of the same units. (The cypress ecotope exhibits only 2,585 units, which might be the effect of nutrient limitation within the swamps.) Florida Bay, however, has a significantly higher developmental capacity (18,540 gCarbon-bits/m2/y vs. 15,744 for the mangroves.) Almost all of this difference can be accounted for by the higher ascendency of the Florida Bay marine system (7,003 gCarbon-bits/m2/y against only 5,608 units in the mangrove.)
The story in words that these numbers are revealing is that Florida Bay appears to be the least stressed of the three communities studied thus far. (This despite the fact that some species displacements have occurred in the Bay in recent years.) The mangrove system is more stressed than the Bay community due to the (natural) variations in osmolarity common to estuaries. The cypress swamps appear to be limited in comparison to both other systems by a dearth of nutrients (probably phosphorus), which are abundant in the marine and estuarine waters and sediments. Some will point out that it may seem incongruous to say that Florida Bay is the least stressed of the three systems, given the serious species displacements that have occurred there in recent years. The other two systems seem almost pristine by comparison. But one must keep in mind the order of magnitude of the disturbances. Whatever stresses are afflicting Florida Bay have affected only the mix of species in that marine community. Its fundamental system processes remain largely intact. The natural stressors affecting the mangroves and the swamp are apparently much greater in magnitude, for they modulate the very rates of material and energy processing. (One should not conclude from this that humans have done little detriment to Florida Bay. Rather, it should be sobering to contemplate how future damage to the ecosystem could be even more catastrophic by comparison with what has already transpired.)
Information indices are usually applied only to whole systems. Evidence is accumulating, however, that the various sub-components of the ascendency-like variables can serve to gauge the contributions of individual system elements to the performance of the whole system (Ulanowicz and Baird 1999). For example, the ascendency is comprised of terms that are generated by each transfer in the system. If, for example, one sums up all the terms generated by the inputs to a given taxon (say, the jth one), the result is a measure of the contribution of that compartment to the full system ascendency (call it Aj). Because ascendency may be viewed as an indicator of efficient system performance (Ulanowicz 1997), the same partial-sum, Aj, represents the contribution of taxon j to overall system performance. Furthermore, if one then divides Aj by the corresponding throughput for taxon j (call it Tj), the ratio Aj/Tj will represent the contribution per unit of activity of j to the total system ascendency.
Calculating and ranking these "relative contribution coefficients" proves to be a most interesting exercise. When the average trophic levels of the 68 compartments of the cypress wetland ecosystem were calculated, for example, the alligators, snakes and wading birds were seen to feed at trophic levels higher than some other "charismatic megafauna", such as the Florida panther, Bobcat, or Grey Fox. The relative contributions to ascendency by the latter, however, actually outweighed those of the former (Ulanowicz et al. 1998). The relative values of these coefficients seem to accord well with most people's normative judgments of the specific "value" of the various taxa to the organization of the system as a whole (Table 7).
The relative contributions of the taxa of the mangrove ecosystem closely mirrored the results obtained from the cypress ecosystem (Ulanowicz et al. 1997) The feline predators (Florida panther and bobcats, #88) ranked only 25 and 27 in terms of trophic level (average levels 3.33 and 3.26, respectively.) Their relative contributions to the system ascendencies were highest of all taxa during both seasons, however. They were followed, as in the cypress glades, by the foxes. Barracuda and dolphins ranked in third and fourth places, just ahead of the snakes. (The snakes in the cypress ecosystem ranked 12th and 7th during the wet and dry seasons, respectively.)
The system indices in the mangrove ecosystem show remarkable little change between wet and dry seasons -- even less than what was observed in the other two biomes.
With 94 compartments in the trophic flow network the number of potential pathways for recycling carbon is roughly proportional to 94-factorial -- an immense number! The fact that the network is not fully connected reduces the number of potential cycles considerably. Nonetheless, the number of simple cycles in the mangrove network remains enormous. Counting them using the standard algorithm in NETWRK 4.2a was broken off after registering ca. 3.3 billion cycles. The situation is similar to what happened last year with the Florida Bay network, and the same approach taken then was followed with the mangrove systems: Most of the overwhelming number of simple cycles counted above involve more than one detrital component. We make the approximation, then, that one may choose to ignore those cycles that contain more than one nonliving compartment. The identification of such "single-detritus" cycles can be achieved first by removing the cycles that contain no detrital links, and then by successively adding the detrital compartments into the search, one at a time. The number of cycles counted in this manner will be a radical underestimate of the total number of cycles present, but once they have been extracted from the network, the residual graph will contain no cycles. Even this simplification takes considerable computing power to execute. The dry season apparently contains fewer routes for recycle (unlike with Florida Bay, where the wet season showed a greater number of cycles), and the abbreviated algorithm took approximately 10 days to execute on a SPARC-Ultra. As of the time of this report, the analysis of the wet season is still underway after more than 2 weeks of runtime. Fortunately, the results from the dry season analysis probably give us a reasonable picture of what is going on in the mangrove ecosystem as concerns the recycling of carbon. The first stage in the cycle analysis was the removal of the cycles that contain no nonliving compartments. These are generally rare in most ecosystems (Pimm 1982). There were only 12 such cycles during the dry season in the mangroves. The cycle of greatest magnitude was the mutual predation of turtles and crocodiles (Large turtles can eat juvenile crocodiles, but the amount cycled over this route is a miniscule 211 micrograms carbon/ m2/ y.
By far and away the greatest number of single- detrital cycles was generated by compartment #92, carbon in the sediment (9,482,523 to be exact.) Of this huge number of cycles, only 15 accounted for 97.5% of all recycling by the system. The major route for recycle was from carbon in the sediment (#92) to sediment bacteria (#8) to meiofauna (#11) and back to the carbon sediment. Secondary routes involved detours through the flagellae (#9) and ciliates (#11) in the sediment, as shown in Figure 9. In this regard, the mangrove estuary is behaving much like conventional estuaries, such as the Chesapeake, where the bulk of recycling activity occurs among the small bodied denizens of the sediment (Baird and Ulanowicz 1989).
Only 59,819 simple cycles were generated by the water column POC alone. The largest of these involved immediate exchange between POC (#93) and free microbes in the water column (#6). This route cycled a mere 5.66 gC/m2/y. Similarly, only one cycle was generated by the dissolved organic carbon in the water column (#94). Along with the POC and free microbes, the DOC cycled 7.96 gC/m2/y.
The aggregate activity devoted
to cycling in the mangrove ecosystem during the dry season is 575.5 gC/m2/y,
which puts the Finn Cycling Index at 17.2%. This value is intermediate to an
FCI of 21% for Chesapeake Bay and 14.4% for Florida Bay, but significantly greater
than the relatively sparse cycling (8 - 9%) occurring inland in the cypress
swamps.
Previous
Page | Next Page | Table
of Contents | Return
to Mangroves