Conceptual and numerical models of the evolution of pedogenic carbonates in soils of arid and semi-arid regions: A review

Introduction Calcareous soils are widely distributed in arid and semi-arid regions of the world and the presence of carbonates in the soils affects both physicochemical properties and the pedogenic evolution. In addition, soil carbon plays a critical role in the global carbon cycle, and pedogenic carbonates are an important sink for atmospheric carbon. Pedogenic carbonates are also the proxy of past climatic variations which are frequently used in geomorphological and paleoclimatological studies. The aim of this paper was to provide a comprehensive review for formation and accumulation of the pedogenic carbonates in arid and semi-arid regions.   Methodology We reviewed the published studies on carbonates formation and accumulation in the soils. Dynamic and processes of carbonate accumulation were explained. Accordin to the literatures, 15 conceptual modeles and 15 numerical and computer modeles of pedogenic carbonares formation were identified.      Results Numerous models and classification systems have been proposed for the formation and evolution of pedogenic carbonates, including conceptual models and numerical and computer models. Indeed, conceptual models are qualitative descriptions of events and changes in the systems, while numerical models are quantitative descriptions of events and processes and can be very challenging due to the very complex nature of the soil system. Generally, conceptual models of calcrete development can be divided into two categories of conceptual models without considering landscape processes (pedon scale models) and conceptual models with considering landscape processes (landscape scale models). At the pedon scale, three types of conceptual models have been defined for the development of calcic horizons: (1) the model of progressive changes, (2) the model of in-situ alteration and (3) the rhizogenic model. On the other hand, five conceptual models have been proposed for the evolution of calcareous soils with considering the processes of erosion, sedimentation and degradation of calcretes. Numerical and computer models of carbonate accumulation in soils can be used to study the influences of late Quaternary climatic changes on the carbonate accumulation in soils. In general, mathematical, compartment and hydrological models have applied in the modeling of calcite accumulation in soils. However, although theses models may be internally correct, they may represent incomplete phenomena due to the lack of enough knowledge of the system or incorrect assumptions.   Discussion In the conceptual models, the process of carbonate formation is considered as crystallization of calcite from soil solution, calcification of plant tissue or recrystallization of calcite crystals in the parent rock. Because conceptual models study the evolution of carbonates over time, chronosequences (e.g. the geomorphic surfaces of alluvial fans) have been used as a tool in the development of these models. Overall, despite the widespread application of conceptual models at the pedon scale by geoscientists, there are some disadvantages:   (1) Many of these models do not consider the role of landscape processes such as erosion and sedimentation in the process of carbonate accumulation. (2) Conceptual models assume the development calcic horizons as evolution from one stage to another in the range of 4 to 8 major evolutionary stages with a linear trend. However, micromorphic analyses have been demonstrated the polygenetic and nonlinear nature of mature calcretes. (3) Many conceptual models have assumptions that may not be appropriate for many parts of the world. Some models consider the presence of interbedded limestone, the presence of clayey sediments or the presence of a natric horizon in the early stages of development. Therefore, these models may have only regional implications. (4) In some models, different formation mechanisms are suggested for similar phenomena. For example the evolution of a laminar layer on the surface of mature calcretes are attributed to the accumulation of moisture and plants roots or the accumulation of aeolian silt by lichens. (5) Recent studies and the application of micromorphological techniques have shown that conceptual models are over simplifications and do not take into account many processes. Despite the issues of conceptual models, these models are still valid and are widely used to describe the morphology of pedogenic carbonates due to their simplicity of application and high efficiency. Computer models have been developed since the early and mid-1980s and have been used primarily to study the effect of Quaternary climatic indicators on the accumulation of carbonates in the soil. Initially, models were developed with an emphasis on physical and chemical processes of calcite deposition in soils and soil system was considered only in the CaCO3-H2O-CO2 system without considering the role of biochemical reactions. Recently, biological factors and the effect of plant roots as well as weathering of aluminosilicates on calcite formation have also been considered. In addition, the assumptions of these modeles to simplify calculations which have a great impact on the results: (1) Rainfall infiltrates into the soil fully and in one dimension and there is no horizontal flow or surface runoff. (2) Soil temperature is assumed to be constant and seasonal variations are not considered. In addition, the use of modern temperature as a long-term variable is incorrect due to Quaternary climatic changes. (3) Water movement through the pedon is assumed as a saturated flow and unsaturated flow, which is important in coarse textured and dry soils, is ignored. In addition, the effects of matric potential changes and upward movement of water in the soil by plant roots, are not considered. (4) Changes in the amount of hydraulic conductivity and soil water holding capacity during the accumulation of carbonates in soils are of important. For example, plugging the soil matrix with carbonates increases the water holding capacity of soils. Furthermore, the formation of carbonate nodules causes changes in the physical properties of the soil, as well as changes in the path and velocity of water in the soil and may result in the preferential dissolution of calcite. (5) In many numerical models, carbonate transport is assumed to be a mass flow and the diffusion of carbonates in the soil is ignored. (6) The origin of calcium ions in many models is considered as a constant rate of dust input and/or a constant concentration of calcium ions in rainwater, which is not reasonable in the long-term due to late Quaternary climate shifts. (7) The amount of present rainfall in many models is considered as long term precipitation, which is not a correct assumption owing to climatic change in the Quaternary. In addition, the models are not able to estimate the effects of rainstorms on the formation of carbonates in arid regions. Studies have shown that heavy rainfalls in arid lands can be very effective in formation of calcic horizons. (8) Information on the partial pressure of carbon dioxide in soils is in many cases inaccurate and also the average annual application of this index is not correct due to seasonal fluctuations and ecological changes during the Quaternary. (9) Many models consider only calcite mineral as soil carbonate mineral and ignore other minerals, especially dolomite. Studies have proven that the secondary calcite that precipitates from the soil solution is richer in magnesium than pure calcite. The mineralogical analysis of pedogenic carbonates of soils demonstrated the occurrence of low-Mg calcite. (10) Numerical models assume that the processes of dissolution and deposition of calcite occur instantaneously and very rapid relative to water movement in soil, which may not be a correct assumption in all soils. (11) Many models do not consider the role of erosion and sedimentation processes in carbonate accumulation and assume that soil thickness is constant during the evolution.   Conclusion Despite the problems of both conceptual numerical models, they are very valuable tools for testing past climatic scenarios on soil development. Finally, a review of conceptual models shows that these models are mainly based on the soils of the arid regions of the southwestern United States, the Mediterranean, and Australia. Although calcareous soils cover most of Iranian lands, there is no modeling for carbonate accumulation processes in these soils. Therefore, there is a very strong potential to test the above models and provide new conceptual models based on the conditions of Iranian soils.