Soil salinity and cotton yield estimation on regional scale in Tarim River Basin using EPIC Model and SOTER approach

Transcription

1 2 nd MEECAL conference: Management of Ecosystems and Environmental Changes in Arid Lands in Central Asia 10 th and 14 th December 2015, Munich, Germany Final Sino-German Conference of SuMaRiO Soil salinity and cotton yield estimation on regional scale in Tarim River Basin using EPIC Model and SOTER approach Hussein Othmanli Karl Stahr Zhao Chengyi 1

2 Problem Description The Tarim Basin is characterized by: 1. Extremely continental desert climate Poor rainfall (50-70 mm a -1 ) High ETp (2,100 3,400 mm a -1 ) 2. Agriculture areas dominated by intensive irrigated cultivation (cotton) Huge water consumption with 60% - 70% water loss Inefficient drainage system (earth channel system) Groundwater level 1-2m below soil surface loamy soil texture consequence: High capillary water movement (loamy texture + high evapotranspiration) Accumulation of salt in the root zones and at the soil surface => Problem of soil salinity 2

3 Problem Description Scientific challenges: Lack of data in the Tarim River Basin Data access due to data policy Own national classification, methodology and publications (in Chinese) Difficult / no access to some areas Rapid land use change Overcome: Own land survey (soil profiles) Remote sensed data (Landsat, SRTM 90m) Prediction methods (some parameters e.g.: FC, K) 3

4 Objectives Characterization of the soils in the area with regard to chemical and physical properties Special task: Assessment and mapping of salinity status Establishment of SOTER-database (SOil & TERrain database) Cotton yield estimation on a regional scale using calibrated and validated EPIC model based on the SOTER database Run several scenarios under various conditions 4

5 Study Area 5

6 I. Characterization of the soils in the region 26 soil profiles in Aksu-Alar region 23 soil profiles in Yingbazar In situ description according to FAO guidelines (2006) Soil classification regarding to WRB (2006 & 2014) Analysis & assessment of the chemical and physical soil properties 6

7 Preparation of soil map I. Characterization of the soils in the region 7

8 Preparation of soil map I. Characterization of the soils in the region 8

9 Preparation of soil map I. Characterization of the soils in the region MLC land cover map- Landsat 8 image Local soil maps Harmonized World Soil Database Soil profile data 9

10 Preparation of soil map I. Characterization of the soils in the region 10

11 (SOil & TERrain Database) is a spatial database with focus on soil and terrain conditions connects various digital maps of different scales with their attribute data forms appropriate input and output data for simulation models on regional scale II. Establishment of SOTER Database Required maps: Geological maps (available) Soil map (prepared) Slope map (from SRTM90m) Land use map? Soil salinity map? 11

12 II. Establishment of SOTER Database Preparing of the land use map Normalized Difference Vegetation Index NDVI > 0.30 NDVI > 0.30 NDVI = (NIR-R) / (NIR+R) Arable Lands Arable Lands 12

13 II. Establishment of SOTER Database Assessment and mapping of salinity status Soil salinity can be estimated by measuring the electrical conductivity of the soil solution/ extraction (1:1), (1:2.5), (1:5), EC (1:2.5) and EC (1:5) done for all soil samples in lab. The unit is Siemens or Mhos ECe: The Electrical Conductivity of Saturated paste: Estimating ECe: ECe = 250*EC(1:2.5) / FC 13

14 ECe (of topsoil) ms/cm II. Establishment of SOTER Database Assessment and mapping of salinity status NDVI vs (ECe ms/cm of topsoil) from 15 soil in arable lands y = x R² = ,40 0,45 0,50 0,55 0,60 NDVI ECe = *NDVI

15 II. Establishment of SOTER Database Assessment and mapping of salinity status ECe = *NDVI

16 II. Establishment of SOTER Database Overlay Geological map Soil map Slope map from (SRTM 90m) Land use map from Landsat 8-Image Soil salinity map 16

17 II. Establishment of SOTER Database 86,000 Pixel = 86,000 ha 17

18 EPIC: Cropping systems model EPIC (Erosion Productivity Impact Calculator) EPIC (Environmental Policy Integrated Climate) for estimation of : crop yield soil erosion C-sequestration III. Cotton yield simulation with the EPIC model on fields scale or areas with homogeneous soils, management and climate. EPIC input data: Soil and terrain data (SOTER-Database) Climate data (T-max, T-min, relative humidity, precipitation, rainy days, wind speed) Crop requirements (e.g. potential heat units) Model run A single standard management scenario applied to the whole area 18

19 Simulated cotton yield [t/ha] Cotton yield [t/ha) Calibration of the EPIC-Model III. Cotton yield simulation with the EPIC model y = x R² = Gemessene Erträge Measured yield Simulated yield Simulierte Erträge Measured cotton yield [t/ha] 0 A01 A03 A05 Comparison of the simulated cotton yield to the measured [t/ha]. Adapted Parameters: Potential heat unit PHU: 1800 C Radiation Use Efficiency RUE: 25 (kg ha -1 MJ -1 m 2 ) 19

20 III. Cotton yield simulation with the EPIC model The total simulated cotton yield in the region under current conditions = 328,700 t 20

21 Average of simulated cotton yield [t/ha] Soil salinity was the major limiting facture for the simulated cotton yield with the EPIC model 7 III. Cotton yield simulation with the EPIC model 6 5 y = e x R² = ECe of topsoil [ms/cm] 21

22 IIII. Run of scenarios Scenarios: Current conditions Total yield: 328,700 t Scenario I Current conditions +: T + 1 C 20% of irrigation water = 400 mm + 2 X TDS (irrigation water) = 1000 mg/l Total yield: 271,870 t Loss 17% Scenario II Current conditions +: T + 2 C 40% of the irrigation water = 300 mm + 3 X TDS (irrigation water) = 1500 mg/l Total yield: 213,960 t Loss 35% n= 86,000 22

23 IIII. Run of scenarios Scenarios: Current conditions Total yield: 328,700 t Scenario I Current conditions +: T + 1 C 20% of irrigation water = 400 mm + 2 X TDS (irrigation water) = 1000 mg/l Total yield: 271,870 t Loss 17% Area = 86,000 ha 23

24 IIII. Run of scenarios Scenarios: Current conditions Total yield: 328,700 t Scenario II Current conditions +: T + 2 C 40% of the irrigation water = 300 mm + 3 X TDS (irrigation water) = 1500 mg/l Total yield: 213,960 t Loss 35% Area = 86,000 ha 24

25 Conclusion The spatial database applied is a useful tool for storage and inquiry of soil and terrain data at various conditions EPIC can be a helpful tool for regional planning and for the decision support system, thus the EPIC model assesses the impact of climate change and management strategies on crop yield production. More calibration data and ground check will enhance the simulation results. For the sustainability of cultivation system in the Tarim River basin several estimations and scenarios of land management, fertilizing and alternative crops should be done. More irrigation water does not mean more yield, hence water saving techniques (deficit irrigation) enhance the water use efficiency and with a proper drainage systems will keep the water table deep under the root zone. 25

26 Research contribution to Ecosystem Services (ESS) in the Tarim Basin and the contribution to the SuMaRiO-Decision Support System (DSS) ESS Estimation of agricultural biomass production DSS Assessment of soil salinity

27 27

28 28

29 measured cotton yield [t/ha] Validation of the EPIC model III. Cotton yield simulation with the EPIC model NDVI vs measured cotton yield [t/ha] on 11 sites ( Sep. 2014) y = 0.006e x R² = NDVI ( ) 0 0,52 0,53 0,54 0,55 0,56 0,57 0,58 0,59 0,60 NDVI Cotton yield (t/ha) = 0.006e (11.996*NDVI) 29

30 III. Cotton yield simulation with the EPIC model Estimated cotton yield from NDVI [t/ha] Cotton yield (t/ha) = 0.006e (11.996*NDVI) 30

31 Simulated cotton yield of EPIC model [t/ha] Validation of the EPIC model III. Cotton yield simulation with the EPIC model Comparison between the simulated cotton yield of EPIC model with the estimated cotton yield from NDVI on 17 arable sites (soil profiles) 8 7 y = x R² = Estimated cotton yield from NDVI [t/ha] 31

32 Validation of the EPIC model III. Cotton yield simulation with the EPIC model Comparison between the simulated cotton yield of EPIC model with the estimated cotton yield from NDVI on 17 arable sites (soil profiles) 18,00 y = 1,8912x R² = 0,8501 Biomass vs Cotton [t/ha] 16,00 14,00 12,00 10,00 8,00 6,00 4,00 2,00 0,00 0,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 8,00 9,00 32

33 Validation of the EPIC model III. Cotton yield simulation with the EPIC model Comparison between the simulated cotton yield of EPIC model with the estimated cotton yield from NDVI on 17 arable sites (soil profiles) 18,00 y = 115,19x - 54,638 R² = 0,8532 NDVI_ / Biomass 16,00 14,00 12,00 10,00 8,00 6,00 4,00 2,00 0,00 0,52 0,53 0,54 0,55 0,56 0,57 0,58 0,59 0,6 33

34 The soil map of Yingbazar region 34