Full text
| |
| Author(s): |
Marina Luciana Abrêu de Melo
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Piracicaba. |
| Institution: | Universidade de São Paulo (USP). Escola Superior de Agricultura Luiz de Queiroz (ESALA/BC) |
| Defense date: | 2020-02-04 |
| Examining board members: |
Quirijn de Jong van Lier;
Simone Toni Ruiz Corrêa;
Fábio Ricardo Marin;
Everton Alves Rodrigues Pinheiro
|
| Advisor: | Quirijn de Jong van Lier |
| Abstract | |
A low soil water availability often results in drought stress, which is the main limitation to sugarcane growth in Brazil. Process-based crop simulation models can help in understanding the factors which determine drought stress. We aimed to estimate the limiting pressure heads of the Feddes transpiration reduction function (h3h, h3l and h4) for sugarcane by inverse modeling with the SWAP/WOFOST model, and to use the calibrated and validated model to simulate crop productivity and water balance components in different soil scenarios. The model was calibrated to reproduce the conditions of a 22-month experiment of sugarcane (2014-2016) in Jaú, São Paulo State (SP), Brazil. The pressure head (h, cm) was measured at 0.2, 0.4 and 1.0 m depths by automated sensors. Soil samples were collected at the same depths to determine the soil water retention curve (SWRC). From the SWRC, h values were converted to soil water content (θ, cm3 cm-3). Actual crop evapotranspiration (ETr, mm d-1) was estimated from data recorded by micrometeorological towers. The Parameter ESTimation program was used to estimate the parameters h3h, h3l and h4 simultaneously from the observed data of θ and ETr. The quality of calibration, before and after the parameters estimation, was assessed by the root mean square error (RMSE), the Nash-Sutcliffe model efficiency coefficient (NSE) and the coefficient of determination (R2). The calibrated model was validated with 30 months (2005-2007) of experimental data of a sugarcane crop grown in Jaboticabal, SP. Determinations of θ and dry matter were used to compare observed and simulated data. The quality of validation was assessed by the same statistical indicators used in calibration. The calibrated and validated model was used to simulate 30 sugarcane cycles (1987-2017), with six soils from the Mesoregion of Piracicaba, SP, Brazil. For each cycle, the crop productivity in dry mass per unit area (kg ha-1) and the components of water balance (transpiration, interception, soil evaporation, runoff and deep drainage) in depths per cycle (mm), representing volume per unit area, were evaluated. In the calibration step, the model simulated satisfactorily θ (RMSE = 0.018 cm3 cm-3, NSE = 0.39, R2 = 0.74) but unsatisfactorily ETr (RMSE = 2.24 mm d-1; NSE = -1.05; R2 = 0.34). The estimated values to represent the limiting hydraulic condition to sugarcane were: h3h, = -460 cm, h3l = -1,968 cm and h4 = -9,646 cm. In the validation step, the model simulated satisfactorily both θ (RMSE = 0.023 cm3 cm-3; NSE = 0.61, R2 = 0.70) and dry mass (RMSE < 7,000 kg ha-1; NSE ≥ 0.85; R2 ≥ 0.90).The soil hydraulic parameters used in the simulations for Piracicaba were not determinant to the crop productivity. Transpiration, soil evaporation, runoff and deep drainage varied significantly with soil conditions. We conclude that the Feddes function implemented in the SWAP/WOFOST model promotes low sensitivity to its parameters, making it difficult to reliably calibrate. Ideally, this study should be conducted from a detailed experiment which considers soil, plant and atmosphere variables measured appropriately. From this ideal experiment, it may be possible to parameterize and calibrate robust physical models which reproduce adequately the complex process of soil water extraction by sugarcane roots. (AU) | |
| FAPESP's process: | 18/12663-8 - Limiting hydraulic condition for drought stress in sugarcane |
| Grantee: | Marina Luciana Abreu de Melo |
| Support Opportunities: | Scholarships in Brazil - Master |