Considerations on the programmed functional life (one generation) of a green artificial reef in terms of the sustainability of the modified ecosystem

Abstract

The installation of artificial reefs serves to enhance marine ecosystems, although it also modifies them. These changes do not have to be irreversible, since it is possible to treat the functional life of an artificial reef (AR) as a variable factor to be determined, with the objective of contributing to the sustainability of the ecosystem. The quest for sustainability does not end with the manufacture and installation of the AR units. It is also necessary to analyse the sustainability of the modified ecosystem, through the production of services. This leads to consider the medium-term return of the ecosystem to its initial state, once the functional life of the ARs expires. This paper presents and justifies an AR design/composition for limited functional life. It is the result of acting on the base material, the concrete, with the objective of limiting the useful life to one social generation. Four different dosages were proposed for such a purpose. They were subjected to mechanical tests (compressive strength and absorption after immersion), including an innovative abrasion-resistant one. The results allow estimating the functional life of the four types of concrete from the design variables (density, compactness, and quantity of water and cement as well as its relation). To this end linear regression models and clustering techniques were applied. The described procedure leads to an AR design for limited functional life.

Keywords: Artificial reef; Design for limited functional life; Linear regression model; Reversible process concept; Sustainability ecosystem.

Graphical abstract

Fig. 1

Changes in the modified ecosystem due to the introduction of a conventional artificial reef (CAR) and a contrived durability green artificial reef (CD – GAR) Source: own based on Thompson [37].

Fig. 2

Evolution of different artificial reef designs in terms of resources and sustainability. OAR: opportunity artificial reef, CAR: conventional artificial reef, GAR: green artificial reef, CD GAR: contrived durability or one generation artificial reef. Source: own.

Fig. 3

Flowchart of the methodology followed in this study to achieve a concrete dosage for artificial reefs with a limited functional life. Source: own.

Fig. 4

Device for the limited durability test. Source: own.

Fig. 5

Concrete samples after the abrasive erosion test: A shows concrete Type 1, B shows concrete Type 2, C shows concrete Type 3 and, D shows concrete Type 4. Panel E shows the profilometer attached to the robot. Source: own.

Fig. 6

Procedure to determine the loss of volume of each concrete sample by using the profilometer and the robot. Source: own.

Fig. 7

Volume loss by double trapezoidal integration. Panel A shows the interpolation of volume below the created plane. Panel B shows the interpolation of volume above the created plane.

Fig. 8

Study of the linear relation between the variables that define the AR concretes. The histogram of each variable is included in the diagonal. The Pearson linear correlation coefficients are included (with p-values of the test defined by $H_0:r=0$), in addition to the scatterplots corresponding to each pair of features with the corresponding linear regression fitting (95% confidence interval is also provided).

Fig. 9

Dependence between critical to quality variables for the concrete and the type of concrete. Panel A shows the influence of the type and age of concrete on the concrete compression strength. In the panel C the dependence between the water absorption and type is observed, whereas the panel B shows the volume loss of concrete depending on the type.

Fig. 10

Panels A–F: Scatterplots with the relation between the concrete critical to quality variables (stress strength, volume loss, water absorbed) and the concrete design variables, that act such as predictors (in the X axis). Panels G–I: Linear relation between the 3 critical to quality variables. Linear regression fittings with 95% prediction confidence intervals, linear regression equations, the p-values of the slope parameters, and the corresponding determination coefficients are also shown.

Fig. 11

Graphical outputs of the clustering analysis performed from the volume loss and water absorption of concrete for AR. Panel A: silhouette plot for detecting the number of clusters. Panel B: dendogram from applying hierarchical clustering with 4 groups. Panel C: graphical output of the k-Means method for 4 groups. Panel D: dendogram from applying hierarchical clustering with 2 groups. E: biplot from the application of the k-Means method for 2 groups using water absorption and volume loss. Panel F: water absorption trends of different types of concrete. Panel D: biplot from the application of the Clara method for 2 groups (from only the water absorption trends).