Modelling aerosol-cloud-precipitation interactions for weather modification in East Africa
Abstract
Precipitation enhancement is one of the possible measures to improve on the available fresh
water in East Africa (EA). The existing gap in knowledge not only on suitable conditions but
also Aerosol-Cloud-Precipitation (ACP) interactions limits the success of precipitation
enhancement through weather modification. This study aimed at modelling the effects of
aerosols on clouds and precipitation for weather modification in EA.
Aerosol and cloud data were retrieved from Moderate Resolution Imaging Spectroradiometer
(MODIS). It comprised of Aerosol Optical Depth (AOD), Fine Mode Fraction (FMF), Cloud Top
Pressure (CTP), Cloud Top Temperature (CTT) and Cloud Optical Depth (COD). Precipitation
data comprised of 3B42 product sourced from Tropical Rainfall Measuring Mission (TRMM),
and comprised of 3B42 product. The Weather Research and Forecasting model (WRF) initial and
boundary conditions utilized the National Centers for Environmental Prediction Final (NCEP
FNL) data. The daily datasets from MODIS and TRMM spanned the period 2001-2012 (12
years). The spatial and temporal analysis utilized the time series and Principal Component
Analysis (PCA). The relationship between aerosol, clouds and precipitation were based Hybrid
Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) and Multiple Linear Regression
(MLR) analysis. Suitable weather modification conditions were identified using Greater Horn of
Africa (GHA) consensus forecast, CAPE and COD and CTP. Evaluation of WRF to simulations
of Deep Convective Clouds (DCCs) was based on updrafts/downdrafts and precipitation.
Aerosols effects on precipitation modification were simulated based on microphysical properties,
precipitation, and convective strength under low, intermediate, and high Cloud Condensation
Nuclei (CCN) and Cloud Droplet Number Concentration (CDNC) scenarios.
In EA, time series (trend and seasonality), Hovmoller and PCA analysis of aerosol, clouds, and
precipitation showed a high spatial and temporal variability. The PCA analysis yielded 13, 20,
11, 9 and 16 Principal Components (PCs) during MAM for AOD, FMF, CTT, OLR and TRMM
3B42 respectively. Similarly, PCA analysis yielded 14, 18, 10, 10 and 17 PCs during OND for
AOD, FMF, CTT, OLR and TRMM 3B42 respectively. The explained variance during both
MAM and OND were all greater than 57%.
The MLR analysis showed that all aerosol and cloud variables with strong factor loading in EA
had a positive relationship with rainfall. The backward trajectory indicated differences in origins
of transported particles in the atmosphere with strong vertical mixing inland with mixed aerosols
due to mountain blocking systems. GHA consensus forecast verification indicated less reliability
due to lower Critical Success Index (CSI) and Heidke Skill score (HSS). However, verified
consensus forecast for MAM 2012 indicated a likelihood of Near Normal (NN) to Below Normal
(BN) rainfall and thus suitable for precipitation enhancement. Areas located centrally to EA
exhibited optimal seedable temperatures of -5oC to -25oC. DCCs dominated pentad 29 (CTP >
440MB, COD>23 and meridional/zonal transition in CAPE >1000J/kg) over Mt. Kenya catchment.
Therefore, Mt. Kenya catchment was selected as representative of seedable conditions in EA.
Evaluation of the efficiency of WRF microphysics shows that Morrison scheme simulated the
initiation of downdraft cumulus core almost at the same time as observed. The initiation of
several updrafts and its associated downdrafts with strong downdrafts below the updraft cores
were comparable to the observed. Accumulated precipitation based on TRMM 3B42 and WRF
model output for Mt. Kenya Region were also found to be comparable to the 24h simulation.
Observed and model simulated the initiation of downdraft Cores was comparable. Microphysical
properties (vertical profiles of mass concentrations of five hydrometeors of cloud water,
rainwater, ice-crystal snow and graupel) showed the complex relationship under three aerosol
scenarios. Precipitation increased with increase in CDNC and CCN from maritime/clean to
continental/polluted conditions and reduced/suppressed at highly polluted conditions
(CDNC>1600cm-3, CCN>2000cm-3). Accumulated total precipitation exhibited a complex
variation (non-linear relationship) under CDNC and CCN scenarios. The mean of core updraft
and maximal vertical velocity increased under intermediate and low CCN scenarios and
decreased under high CCN scenarios. The response of precipitation to increase in aerosol
concentration (CDNC and CCN) was non-monotonic.
The study indicates the possibility of enhancing precipitation in locations with similar conditions
to Mt. Kenya catchment. Increasing the available fresh water in EA will spur sustainable
development. However, critical issues remain yet to be solved and require stronger scientific
evidence/support. These include improvement of the predictability of seasonal rainfall and
development of cloud-resolving models for the region. Further, It will be necessary to develop
relevant policies to address the benefits, risks, and ethical issues related to weather modification.
Publisher
University of Nairobi