3.2. Climate change trend analysis for the period 2000–2020.
After the data quality control process, which was carried out via homogeneity tests, RclimDex software was used to analyze the temperature and precipitation data. However, RclimDex provides additional quality control methods for the data before producing climate indices.
The analysis was conducted based on average climatic parameters (precipitation rates and temperatures) for the stations Amman Airport, Jerash, and Khirebit Es Samra. In this paper, the trend is considered significant if the p value is less than 0.05.
a) Temperature trend analysis
Table 6 presents the temperature extreme indices for Khirebit Es Samra, Jerash, and Amman Airport stations with a p value for each index. Accordingly, the monthly max value of daily mean temperature (TMAXmean), monthly min value of daily mean temperature (TMINmean), tropical nights (TR20) and percentage of days when minimum temperatures fall below the 10th percentile (tn10p) were found to be statistically significant climate change indices.
Warm nights (TR20) were the most significant indicators that showed a significant change at all stations (p < 0.02). The stations showed a positive trend with a slope ranging between 1.03/yr and 2.08 /yr. The results indicated that the number of nights in which temperatures exceeded 20 degrees Celsius during the past 20 years increased by 21 nights for Khirebit Es Samra, 42 nights for Jerash and 32 nights for Amman Airport.
The indices (TXn and TNn) concerned with temperatures during the coldest days and coldest nights show slight changes during the period of 2000–2020. Amman Airport station witnessed the smallest declines (-0.005 and 0.004, respectively). Jerash station witnessed the maximum decline (0.034) with respect to the TXn index, while the maximum decline for TNn was at Khirebit Es Samra station (-0.029).
The cold spell duration index (CSDI) has a negative trend, ranging from 0.08/yr. to 0.3/yr. For the DTR index, which represents the difference between the maximum and minimum temperatures during one day, the results showed a decrease of 0.01/yr. and 0.02/yr.
b) Precipitation trend analysis
Table 7 shows the linear trends of the annual precipitation indices in the study area. Most of the precipitation indices showed decreasing trends except for the CDD index, which indicates an increasing trend in the maximum dry period. As expected, most indicators showed decreasing trends during the study period, indicating a continuous decrease in the rates and intensity of rainfall in the study area. It can be concluded that there are small changes in the total annual precipitation over the study area, which is consistent with a previous study conducted by Al-Qaisi (2010).
Table 6
Annual linear trends of the extreme climate indices of temperature over the study area
Indices | KHIREBIT ES SAMRA | | Jerash | | AMMAN AIRPORT | |
| Slope | p value | Slope | p value | Slope | p value |
TMAXmean | 0.036 | 0.109 | 0.045 | 0.047 | 0.042 | 0.052 |
TMINmean | 0.057 | 0.003 | 0.059 | 0.002 | 0.056 | 0.002 |
su25 | 0.191 | 0.635 | 0.5 | 0.371 | 0.401 | 0.467 |
tr20 | 1.027 | 0.022 | 2.077 | 0.001 | 1.579 | 0.001 |
fd0 | -0.145 | 0.451 | 0.051 | 0.442 | -0.005 | 0.93 |
gsl | -0.142 | 0.496 | 0 | 1 | 0 | 1 |
txx | 0.02 | 0.768 | 0.025 | 0.755 | 0.029 | 0.676 |
txn | 0.008 | 0.91 | 0.034 | 0.569 | -0.005 | 0.937 |
tnx | 0.067 | 0.333 | 0.105 | 0.036 | 0.127 | 0.017 |
tnn | -0.029 | 0.647 | -0.023 | 0.741 | 0.004 | 0.951 |
tx10p | -0.165 | 0.127 | -0.207 | 0.072 | -0.198 | 0.086 |
tx90p | 0.21 | 0.294 | 0.144 | 0.401 | 0.192 | 0.3 |
tn10p | -0.477 | 0.002 | -0.559 | 0.003 | -0.585 | 0 |
tn90p | 0.336 | 0.058 | 0.509 | 0.007 | 0.448 | 0.02 |
wsdi | 0.317 | 0.41 | 0.387 | 0.202 | 0.403 | 0.358 |
csdi | -0.081 | 0.693 | -0.113 | 0.603 | -0.274 | 0.206 |
dtr | -0.021 | 0.069 | -0.014 | 0.306 | -0.014 | 0.213 |
Table 7
Annual linear trends of the extreme climate indices for precipitation over the study area
Indices | KHIREBIT ES SAMRA | Jarash | | AMMAN AIRPORT |
| Slope | p value | Slope | p value | Slope | p value |
rx1day | 0.12 | 0.611 | -0.131 | 0.711 | -0.123 | 0.7 |
rx5day | -0.076 | 0.871 | -0.639 | 0.515 | -0.739 | 0.345 |
sdii | -0.022 | 0.504 | -0.035 | 0.513 | -0.048 | 0.271 |
r10 mm | -0.078 | 0.232 | -0.236 | 0.143 | -0.299 | 0.005 |
r20 mm | 0.029 | 0.314 | -0.004 | 0.955 | -0.049 | 0.248 |
R25mm | 0.003 | 0.874 | 0.047 | 0.316 | 0 | 1 |
cdd | 0.204 | 0.9 | 0.632 | 0.609 | 0.087 | 0.95 |
cwd | 0.025 | 0.505 | -0.082 | 0.177 | -0.043 | 0.408 |
r95p | -0.369 | 0.694 | 0.671 | 0.703 | -1.458 | 0.308 |
r99p | 0.194 | 0.741 | 0.358 | 0.743 | -0.35 | 0.647 |
prcptot | -1.885 | 0.229 | -3.141 | 0.353 | -5.217 | 0.016 |
3.3. Climate Change Downscaling
The daily data of precipitation, maximum and minimum temperatures, and humidity for the period (2000–2020) for the stations Amman Airport, Jerash, and Khirebit Es Samra were used to determine potential climate change in the basin. The data were formatted using Microsoft Excel and Notepad++, and the data for each parameter were arranged in order from 1/1/2000 through 31/12/2020. The data were sorted, and the value 0 was given for nonrainy days.
a) Model Calibration
SDSM software uses two procedural approaches to model calibration based on climatic data. The first method is the conditional process, which is used when there is an indirect link between precipitation data and regional scale predictors. The second method is the unconditional process, which is used for temperature data and is based on the idea that there is a direct link between climate data and forecasts. The historical predictand data are divided into two time periods: the calibration period is from 2000 to 2020, and the validation period is from 2020 to 2022. The R² values for precipitation and maximum and minimum temperatures were 0.9519, 0.9074, and 0.8682, respectively, indicating good model performance.
b) Model Validation
After a successful calibration test, a validation test was conducted for the period 2020–2022 to determine the correctness of the model that would be downscaled for future estimates. The results showed that the model run has received sufficient validation and can be used for future forecasts.
c) Future Climate prediction
The study shows that the rate of precipitation drops will be the lowest in the RCP 2.6 scenario, with a decrease of 12% in 2070 compared to 2020. The drops in the RCP 4.5 and RCP 8.5 scenarios will reach 22% and 34%, respectively. The data of daily maximum and minimum temperatures for the period 2000–2020 were used for the RCP 8.5, RCP 4.5, and RCP 2.6 emission scenarios. The results indicate a rise in the average annual minimum temperature, as the average minimum temperature in 2070 will be 15.13, 15.31, and 15.46°C for RCP 2.6, RCP 4.5, and RCP 8.5, respectively, compared to 14.7°C in 2020. In contrast, during the same period, the average annual maximum temperatures will increase in the RCP8.5 scenario, reaching 29.95°C. Table 8 shows the percentage change in precipitation rates and the maximum and minimum temperatures during the study period of 2020–2070.
Researchers have conducted studies and modeling to predict future climate changes in the Amman-Zarqa basin. The findings of these studies suggest that the mean temperature in the region is expected to rise by 2.1°C (1.7 to 3.1°C) for the Representative Concentration Pathway (RCP) of 4.5 and 4°C (3.8 to 5.1°C) for RCP 8.5 with warmer summers. These predictions were made using downscaled regional circulation models and the Soil Water Assessment Tool (SWAT) (Al-Hasani et al. 2023). Another study conducted by Al-Karablieh (2011) in the Zarqa River basin predicted a decrease in rainfall amount by 15–30% due to the direct and indirect impacts of climate change scenarios.
Table 8
Annual decline rate of climate parameters for RCP 2.6, RCP 4.5, and RCP 8.5
parameter | scenario | 2020 | 2030 | 2040 | 2050 | 2060 | 2070 |
Total rainfall | | | | | | |
| RCP2.6 | 0 | 2 | 6 | 8 | 10 | 12 |
| Rcp4.5 | 0 | 4 | 8 | 12 | 17 | 22 |
| Rcp8.5 | 0 | 3 | 8 | 16 | 24 | 34 |
Maximum temperature | | | | | | |
| RCP2.6 | 0 | 0.904 | 1.163 | 1.574 | 1.746 | 1.762 |
| Rcp4.5 | 0 | 1.023 | 1.388 | 1.702 | 1.75 | 1.998 |
| Rcp8.5 | 0 | 1.55 | 1.596 | 1.742 | 1.917 | 2.128 |
Minimum temperature | | | | | | |
| RCP2.6 | 0 | 1.56 | 2.072 | 2.498 | 2.579 | 2.606 |
| Rcp4.5 | 0 | 1.845 | 2.616 | 3.222 | 3.377 | 3.701 |
| Rcp8.5 | 0 | 1.869 | 2.637 | 3.491 | 3.953 | 4.667 |
3.4. Groundwater Numerical Modeling
a) Steady state simulation
Steady-state calibration involves matching the observed heads of the aquifer from monitoring wells with the hydraulic heads simulated by the MODFLOW model for a specific year (2020). This is done by sequentially altering the model parameters, such as hydraulic conductivity and vertical anisotropy. The calibration is based on the values of the heads observed by the monitoring wells in the study area. Initial values for vertical anisotropy and hydraulic conductivity are estimated from previous studies related to the Amman-Zarqa Basin.
After the first simulation, the locations of the observation wells and the head values for each well are added to the model. The run option is then converted to parameter estimation mode. The calibration is carried out using the trial-and-error method and creating a scatter plot of points that are randomly distributed based on the well locations and boundary conditions. The goal of the calibration is to reach the best values for hydraulic conductivity. The horizontal hydraulic conductivity values vary in the range of 11.9 m/d (1.38 x 10 − 4 m/s) to approximately 80.2 m/d (9.27 x 10 − 4 m/s). Due to the heterogeneity and complexity of the aquifer, it is not possible to establish trends for the distribution of hydraulic conductivity values in the study area. The simulation results based on calibrated hydraulic conductivity values indicate a high agreement between the modeled hydraulic heads and the observed ones on-site using monitoring wells.
The verification was performed based on the maximum allowable error of 1 meter, and the verification indicates that the difference between the observed and computed values is smaller than the maximum allowable error, as shown in Table 9.
Table 9
Measures of fit between observed and computed heads for steady state (calibration stage)
ID | Palestine Grid North (PGN) | Palestine Grid East (PGE) | Observation Head | Observation Head interval (m) | Observation Head conf (%) | Obs. Head std. dev | Computed Head | Residual Head |
AL1813 | 1152556 | 241006 | 702.2 | 1 | 95 | 0.5102 | 702.207 | -0.0065 |
Al2690 | 1154225 | 242100 | 692.17 | 1 | 95 | 0.5102 | 692.174 | -0.0036 |
AL1444 | 1158318 | 245970 | 594.85 | 1 | 95 | 0.5102 | 594.854 | -0.0035 |
b) Transient simulation (time-dependent simulation)
A simulation of the transient state was conducted to model the hydraulic system (hydraulic head) in a study area. The simulation was divided into two parts: a calibration phase from 1985–2009 and a validation phase from 2010–2015. During the calibration phase, the pumping rates were adjusted as the only changing condition in the simulation. The aim was to match the simulated hydraulic heads in the monitoring wells with the observed data for the corresponding period. This calibration process helps ensure that the simulation accurately represents the behavior of the groundwater system. The steady-state calculations were used as input data for the transient simulation. The transient simulation aimed to simulate the current drawdown and predict the long-term state of the hydraulic system. The simulation started in 1985 when water pumping began (with a pumping rate of 135 MCM) and continued until 2009 (with a pumping rate of 158 MCM). The 25-year period was divided into 25 stress periods to cover the pumping period from 1985–2009.
The following figure shows the modeled and observed head values based on the PEST calculation for AL1444.
c) Model Validation
The validation process involved simulating the model from 2010 to 2015 and comparing the results with the observed values for the same period. The figure below shows the observed and modeled head values for three monitoring wells.
d) Model prediction
Model predictions were used to assess the impact of climate change on the hydraulic head response. Two future pumping rate scenarios were considered. The first scenario assumed a constant pumping rate of 21.8 MCM/year from 2020 to 2070. The second scenario assumed a 20% increase in the pumping rate every 10 years, resulting in a doubling of the pumping rate by 2070. The study period of 50 years was divided into 5 stress periods and 10 time steps for each period, resulting in a total of 50 time steps. This allowed for the examination of the results for each year during the study period.
Two different scenarios were created based on pumping rates. Each scenario includes subscenarios related to emissions rates (RCP 2.6, RCP 4.5, and RCP 8.5) to examine the impact of climate change on groundwater, as well as a scenario that assumes no future climate changes. The explanation of the scenarios is provided below.
This scenario includes three different cases based on three climatic scenarios that include a reduction in the recharge values at different rates (RCP 2.6, RCP 4.5 and RCP 8.5).
a) Climate scenario (RCP 2.6)
In this scenario, recharge will decrease by 2%, 6%, 8%, 10%, and 12% for 2030, 2040, 2050, 2060, and 2070, respectively.
The maximum drawdown for this case was 24.5 meters in 2070.
b) Climate scenario (RCP 4.5)
In this scenario, recharge will decrease by 4%, 8%, 12%, 17%, and 22% for 2030, 2040, 2050, 2060, and 2070, respectively.
The maximum drawdown for this case was 42.5 meters in 2070.
c) Climate scenario (RCP 8.5)
The current recharge rate of 0.02 m/y will be reduced by 3%, 8%, 16%, 24%, and 34% for 2030, 2040, 2050, 2060, and 2070, respectively. This scenario was chosen to include the emission maximums.
The maximum drawdown for this case was 62.7 meters in 2070.
The following table shows detailed values of the modeling results. According to the first scenario, the maximum expected drop value reached approximately 63 meters in 2070 according to RCP 8.5 due to the decrease in the recharge value to its lowest value during the study period. The results also show a convergence in the level of drop in the period 2030 and 2050 in RCP 4.5 and RCP 8.5. The reason for this is that RCP 4.5 includes emissions reaching their peak during the middle of the century and then starting to decrease until they stabilize at 4.5 (W/m2) in 2100, while RCP 8.5 includes an upward increase until it stabilizes at 8.5 (W/m2), which means convergence of emissions values between the two scenarios in the middle of the century.
Table 10
Dataset info for the first scenario for head decline.
| RCP 2.6 | RCP 4.5 | RCP 8.5 |
Date | 1/1/2030 | 1/1/2050 | 1/1/2070 | 1/1/2030 | 1/1/2050 | 1/1/2070 | 1/1/2030 | 1/1/2050 | 1/1/2070 |
Min value (m) | 0.4 | 5.1 | 12.5 | 0.8 | 7.6 | 20.3 | 0.6 | 7.9 | 27.1 |
Max value (m) | 1.5 | 14.1 | 24.5 | 3.0 | 20.2 | 42.5 | 2.2 | 23.6 | 62.7 |
Range (m) | 1.1 | 9.0 | 11.9 | 2.2 | 12.6 | 22.2 | 1.7 | 15.6 | 35.6 |
Mean (m) | 1.0 | 10.4 | 19.7 | 2.0 | 15.1 | 33.5 | 1.5 | 17.2 | 48.2 |
Median (m) | 1.3 | 12.4 | 21.4 | 2.6 | 17.7 | 37.1 | 1.9 | 20.7 | 54.6 |
An increase in the pumping rate in the study area was assumed due to the increase in population growth in addition to the waves of refugees, which led to an unprecedented increase in the water demand. This scenario includes doubling the current pumping rate in 2070.
This scenario was simulated according to four cases. Three of them are the same as in the first scenario, while the last case assumes that there are no future climate changes.
a) No future climate changes
In this case, the current recharge value of 0.02 m/y was adopted with increasing pumping rates during the study period.
The maximum drawdown for this case was 3 meters in 2070.
b) Climate scenario (RCP 2.6)
This case is the same as in the first scenario, with the same rates of reduction but with an increase in pumping rates. As expected, the predicted values for the decrease in the groundwater level were greater in this scenario due to the higher withdrawal rates according to the same expected climate, where the maximum drawdown for this case reached 27.3 meters in 2070.
c) Climate scenario (RCP 4.5)
This condition was simulated according to the same reduction values related to recharge in the first scenario for the same situation but with the addition of an increase in pumping rates.
This case showed greater decrease values compared to the same case in the first scenario due to the effect of increased pumping on the groundwater quantities, which leads to a greater depletion of the aquifer.
d) Climate scenario (RCP 8.5)
This case is considered the most severe case, as it includes the largest reduction in recharging values, in addition to the increasing pumping rates during the study period.
This scenario, according to this case, witnesses the highest rates of decline in groundwater levels compared to all previous cases in both scenarios. The maximum drop according to this scenario was in 2070 by approximately 65 meters.
Table 4.11 shows detailed values for the modeling results. According to the second scenario, the maximum expected drop value was approximately 66 meters in 2070 according to RCP 8.5. This difference was due to the increase in the pumping rate in this scenario during the study period. The results also show convergence in the level of decline in 2030 and 2050 in RCP 4.5 and RCP 8.5 for the same reason that was mentioned in the first scenario.
The study results indicate that the current pumping practices in the study area are unsustainable, especially because of climate change effects, particularly on the recharge values. The simulation results showed that the groundwater levels in the study area could drop dramatically at a rate of more than 1 m/y according to the worst scenario, which is the second scenario that includes climate changes according to RCP 8.5, while the rate of decline in the first scenario, according to the same case, was 0.95 m/y. year. The rate of decline was more than 0.7 m/y according to the second scenario with RCP 4.5, while it was 0.65 m/y according to the first scenario. Regarding RCP 2.6, the decline rate according to the first scenario was 0.4 m/y, while it increased to 0.45 m/y according to the second scenario.
The impact of climate change on groundwater level drawdown in the Amman Zarqa Basin has been extensively studied and evaluated by researchers. Climate change has been identified as one of the factors that significantly affects water resources in the basin. According to a study by Al-Qaisi in 2010, the yearly drawdown in groundwater level in the Amman Zarqa Basin is estimated to be approximately 0.5 meters. Another study by Al-Shibli in 2018 projected that groundwater levels in the basin may experience a drawdown of approximately − 0.51 meters per year over the next 35 years. These findings suggest that climate change is likely to contribute to the decline in groundwater levels in the Amman Zarqa Basin.
The combination of increased temperatures and decreased precipitation due to climate change can lead to increased water demand and reduced recharge of groundwater. This, in turn, results in a decline in groundwater levels over time (Al-Shibli, 2018).
Table 13
Dataset info for the second scenario for head decline.
| No future climate changes. | RCP 2.6 | RCP 4.5 | RCP 8.5 |
Date | ---- | 1/1/2050 | 1/1/2070 | 1/1/2030 | 1/1/2050 | 1/1/2070 | 1/1/2030 | 1/1/2050 | 1/1/2070 | 1/1/2030 | 1/1/2050 | 1/1/2070 |
Min value (m) | ---- | 0.1 | 0.7 | 0.2 | 5.2 | 13.2 | 0.6 | 7.7 | 21.0 | 0.4 | 8.1 | 27.8 |
Max value (m) | ---- | 1.3 | 3.0 | 1.3 | 15.2 | 27.3 | 2.7 | 21.2 | 45.2 | 2.0 | 24.6 | 65.3 |
Range (m) | ---- | 1.2 | 2.3 | 1.1 | 10.0 | 14.1 | 2.1 | 13.5 | 24.3 | 1.6 | 16.6 | 37.6 |
Mean (m) | ---- | 0.6 | 1.8 | 0.9 | 11.0 | 21.4 | 1.9 | 15.7 | 35.3 | 1.4 | 17.8 | 50.0 |
Median (m) | ---- | 0.6 | 1.6 | 1.1 | 13.0 | 23.0 | 2.5 | 18.4 | 38.7 | 1.8 | 21.4 | 56.1 |