Modeling soil water flow and retention on pseudogley soil in sloping vineyard

Pseudoglej je drugo najzastupljenije tlo unutar agroekosustava, a karakterizira ga slabije propusni potpovršinski horizont koji utječe na periodičko stagniranje oborinske vode i restrikcije u rastu korijena. S obzirom na prisutne klimatske promjene, izrazito je važno pratiti režim vlažnosti tla kako bi se na vrijeme reagiralo i umanjile štetne posljedice. Cilj ove disertacije bio je utvrditi svojstva rigolanog tla iz pesudogleja obronačnog u uvjetima uzgoja vinograda te utjecaj svojstava tla i proraštenosti tla korijenom na tok i retenciju vode u tlu. TDR senzori za praćenje režima vlažnosti postavljeni su na pet pozicija obronka u površinski (Ap) i potpovršinski (Bg) horizont te u red i međuredni dio vinograda. Na istim pozicijama utvrđena su fizikalna, kemijska i hidraulička svojstva tla. Rezultati analiza fizikalnih svojstava tla pokazuju izrazitu heterogenost s obzirom na položaj na obronku. Tako je utvrđen signifikantan utjecaj pozicije na sadržaj pijeska, praha, gline, humusa, volumne gustoće, kapaciteta tla za vodu, kapaciteta tla za zrak, i ukupnog poroziteta (za sve P < 0,0001). Isto tako utvrđene su značajne razlike u fizikalnim svojstvima između reda i međurednog prostora vinograda, za oba horizonta. Tijekom istraživanja uočene su i značajne razlike u volumetrijskom sadržaju vlage u tlu ovisno o poziciji na obronku (P < 0,0001) te između reda i međurednog prostora vinograda (P < 0,0001). Prosječno najveće vrijednosti volumetrijskog sadržaja vlage u tlu zabilježene su na višim pozicijama obronka (tj., P1 i P2). Navedeno bi moglo biti povezano sa povećanim sadržajem gline na istim pozicijama. Razlike u sadržaju vlage između reda i međurednog prostora pripisane su razlikama u korijenu kultura (vinova loza i trava) koje koriste različite količine vode tijekom vegetacije. HYDRUS 1D model korišten za simuliranje toka i retencije vode na istraživanoj parceli kalibriran je na temelju vrijednosti o volumetrijskom sadržaju vlage u tlu izmjerenim pomoću TDR senzora. Hidraulički parametri korišteni kao ulazni podaci u modelu prilikom simulacija procijenjeni su na tri načina, i to temeljem (i) evapotranspiracijske metode pomoću HYPROP i WP4C uređaja, (ii) „tipa pet“ pedotransfernih funkcija u ROSETTA softveru, te (iii) preko inverznog modeliranja u HYDRUS 1D modelu. Neke simulacije sa hidrauličkim parametrima iz HYPROP i ROSETTA softvera bile su nestabilne (model nije konvergirao). Korištenjem inverznog pristupa, sve simulacije bile su stabilne i prokazale su relativno zadovoljavajuće rezultate (RMSE = 0,004 do 0,084 cm3 cm-3). Hidraulički parametri tla procijenjeni inverznim modeliranjem dalje su korišteni kao ulazni parametri za simulacije u HYDRUS 2D/3D modelu kojim se simulirao tok i retencija vode na cijelom istraživanom obronku, te u tri izmijenjena klimatska scenarija. Za još uspješnije korištenje modela, preporuča se posebnu pažnju posvetiti ulaznim parametrima kao što su hidrauličke značajke tla i parametri uzgajane kulture.Pseudogley is the second most widespread soil found in the agroecosystems of Croatia (307,453.2 ha). It is characterized by the presence of a less permeable subsurface horizon that affects the occasional (longer or shorter) occurrence of excessive saturation of the upper zone of the profile and have restrictions on root growth. According to FAO (2014), the soil profile of the natural pseudogley profile is designated as A-Eg-Bg-C, while according to the Soil Classification of Croatia as A-E / S-B / S-C (Husnjak, 2014). Since pseudogley as a soil type develops in regions that are in terms of relief and climate suitable for agriculture, a significant part of pseudogley is deeply ploughed for growing permanent crops. Deep ploughing resulted in mixing of the humus-accumulative with the eluvial-pseudogley and the upper part of the illuvial-pseudogley horizon, thus forming an anthropogenic horizon. Such pseudogleys are referred to as Ap-Bg-Cg (FAO, 2014), or P-B / S-C / S (Husnjak, 2014). Soil moisture distribution and its spatial and temporal variability affect plant growth and development. Due to climate change, it is very important to develop models that will predict their impact on soil, but also on crop production. Given the above, it is extremely important to continuously monitor the water content in the soil with the aim of irrigation scheduling in order to mitigate the effects of drought. Measuring the soil moisture regime requires a lot of field work, is time consuming and expensive. With the development of computer technology, many numerical models have been developed that, based on pedotransfer functions and/or measured hydraulic parameters, can solve water movement in the (un)saturated zone of the soil. Based on the above stated, the following research hypotheses were set: (i) in the investigated agroecological conditions during the autumn/winter period there is no occurrence of prolonged excessive soil saturation (over 120 days per year) characteristic for pseudogley, (ii) soil moisture regime and water flows are affected by relief or position on the slope and hydraulic parameters of the subsurface horizon, and vine roots, (iii) using a numerical model one can reliably determine the flow and retention of water in the studied agroecological conditions. The research was conducted over two years (2019, 2020) in Jastrebarsko (45º41'22 "N; 15º38'22" E). The investigated plot is located on the slope with the southeastern exposure, it is 90 m long and has a slope of 14%. A vineyard (Traminac cv.) was planted on the investigated plot in 1999, where the planting distance in the row was 1.0 m, and between the rows 2.5 m. The rows are oriented down the slope, and the interrow area is planted with grass to limit the erosion. On the investigated plot, five pedological profiles were dug, to a depth of 110 cm, which were equidistant from each other (18 m) along the entire length of the slope. From the pedological profiles, disturbed samples and undisturbed soil cores were taken (100 and 250 cm3 volume cylinders). Samples were taken in a row and in the inter-row part of the vineyard. After sampling, TDR sensors were installed in each of the five profiles for monitoring of soil moisture regime at 30 and 90 cm deep in a row and in the inter-row part of the vineyard. Also, at 30 cm in a row and in the inter-row space, tensiometers were placed to monitor soil water potential, after which the profiles were buried again. In addition to monitoring the soil moisture regime (TDR sensors and/or tensiometers), the level of water stress in the vine can be determined by measuring the leaf water potential of the vine. Therefore, during 2020, at three slope positions (top, middle and bottom of the slope), the water potential on the vine was measured in order to determine the correlation between the soil water potential and the vine (leaf) water potential. The results showed that no significant correlation was observed between the water potential of the soil and the vine. However, the correlation between soil water potential and relative humidity was positive and strong (r = 0.56 - 0.62). Based on standard laboratory analyzes, the basic physical (soil texture, bulk density, differential and total porosity), chemical (pH and humus content) and hydraulic properties of the soil (θs, θr, α, n, Ks) were determined. The results of the analyzes indicate a pronounced heterogeneity of particle size distribution, bulk density, total porosity with respect to the position on the slope, and pH and humus (for all values of P < 0.0001). Significant differences in soil properties between row and inter-row space of vineyards were also found. During the research, significant differences in soil moisture were observed depending on the position on the slope (P < 0.0001), as well as differences between the row and the inter-row space of the vineyard (P < 0.0001). Thus, the highest values of soil moisture content during the investigated period were observed at the upper P1 and P2 positions of the slope. This could be related with increased clay content at these positions. On average, the highest moisture content in the soil during the investigated period was observed in the row for the surface (Ap) and in the interrow for the subsurface (Bg) horizon. TDR sensors determined that excessive saturation was present in the middle of the slope (P3 position) throughout the study period. This could be related to increased clay content, at P1 and P2 slope positions which could affect the occurrence of subsurface lateral flow. However, although saturation conditions prevailed at the P3 position throughout the investigated period, the relative water content in the soil was lower compared to the positions above. This could be related to the increased sand content in the Bg horizon at P3 compared to other positions and to the fact that the sand has a low water retention capacity. In addition to measuring the soil moisture regime, the HYDRUS 1D and HYDRUS2D / 3D model was used for its assessment. The hydraulic parameters used in the model were estimated in three ways: (i) by evapotranspiration method based on HYPROP and WP4C systems, (ii) by “type five” estimation of pedotransfer functions in ROSETTA software, and (iii) by inverse simulation using HYDRUS 1D program. In the first step, the hydraulic parameters (θs, θr, α, Ks, n and l) of the soil obtained through the HYPROP and WP4C systems were used, by applying which the HYDRUS 1D model was unstable, i.e., did not converge in most cases. In the second step, the estimated hydraulic parameters also did not give satisfactory results in HYDRUS 1D simulations, so in the third step they were estimated by inverse modeling. Inverse modeling was performed on the calibration period (from 13.9. to 12.10.2019) Which was chosen because it had the equal time of wet and dry days. Values of R2 were generally above 0.90 excluding P1 and P3 positions in a row in which R2 was, respectively, R2 = 0.645 and R2 = 0.882. RMSE values in most positions are generally less than 0.03 cm3 cm-3, except for P1 position in the row where the RMSE was 0.05 cm3 cm-3. The inverse simulation showed the most reliable (but still not ideal) results of the assessment of hydraulic parameters. However, in the absence of insufficiently good measurements (HYPROP) and estimation of hydraulic parameters (ROSETTA), hydraulic parameters estimated through inverse modeling were further used in the work. After calibration and validation in the HYDRUS 1D program, the hydraulic parameters estimated by inverse modeling were further used as input parameters for simulations of water flow and retention on the entire slope using the HYDRUS 2D/3D program. With respect to climate change in the HYDRUS 2D/3D program, three climate scenarios were simulated (according to IPCC, 2014) (i). mild (predicting a 1.5ºC temperature rise) (ii) medium predicting a 2.5ºC temperature rise (iii) extreme predicting a 3.5ºC temperature rise. Also, IPCC, 2014 predicts that precipitation in the winter months will increase, while longer periods of drought will prevail in the summer months. Although no significant differences were found between the simulated climate scenarios, it can be concluded that the HYDRUS model after calibration and validation can be successfully used to predict different climate scenarios. For more successful use of the model and based on the research results obtained in this dissertation, it is recommended during modeling to pay attention to proper soil hydraulic properties determination, since they have the largest impact on the simulation results. Furthermore, for simulations over longer time periods, the hydraulic properties of the soil should be determined on several occasions with respect to their temporal variability which is pronounced in intensively managed arable soils. In addition to the hydraulic properties of the soil, the data on the crop (eng. leaf area index, root depth) have a significant impact on the simulation results and should therefore be given additional attention