def test_soil_moisture_potential(self):
eps = 0.01
nhru = 100
nsoil = 5
model = richards.richards(nhru, nsoil)
#define parameters
model.theta[:] = np.random.uniform(low=0.05, high=0.5,
size=nhru)[:, np.newaxis]
model.thetar[:] = 0.1
model.thetas[:] = 0.5
model.b[:] = np.random.uniform(low=3, high=8, size=nhru)
model.satpsi[:] = 0.1
#choose layer
psi = []
bpsi = []
for il in xrange(nsoil):
psi.append(model.calculate_soil_moisture_potential(il))
tmp = model.theta[:, il]
m = (tmp <= (1 + eps) * model.thetar)
tmp[m] = (1 + eps) * model.thetar[m]
bpsi.append(model.satpsi *
((model.theta[:, il] - model.thetar) /
(model.thetas - model.thetar))**(-model.b))
psi = np.array(psi)
bpsi = np.array(bpsi)
self.assertTrue(np.allclose(psi, bpsi))
def initialize_richards(self, ):
from pyRichards import richards
#Initialize richards
self.richards = richards.richards(self.nhru, self.nsoil)
#Set other parameters
self.richards.dx = self.dx
self.richards.nhru = self.nhru
#print(self.nhru)
self.richards.m[:] = self.input_fp.groups['parameters'].variables[
'm'][:]
#self.richards.dem[:] = self.input_fp.groups['parameters'].variables['dem'][:]
self.richards.dem[:] = self.input_fp.groups['parameters'].variables[
'hand'][:]
self.richards.slope[:] = self.input_fp.groups['parameters'].variables[
'slope'][:]
#self.richards.hand[:] = self.input_fp.groups['parameters'].variables['hand'][:]
self.richards.area[:] = self.input_fp.groups['parameters'].variables[
'area'][:]
self.richards.width = sparse.csr_matrix(
(self.input_fp.groups['wmatrix'].variables['data'][:],
self.input_fp.groups['wmatrix'].variables['indices'][:],
self.input_fp.groups['wmatrix'].variables['indptr'][:]),
shape=(self.nhru, self.nhru),
dtype=np.float64)
#dtype=np.float64)[0:self.nhru,0:self.nhru]
#print(self.richards.width.shape)
self.richards.I = self.richards.width.copy()
self.richards.I[self.richards.I != 0] = 1
return
def initialize_richards(self,):
from pyRichards import richards
#Initialize richards
self.richards = richards.richards(self.nhru,self.nsoil)
#Set other parameters
self.richards.dx = self.dx
self.richards.nhru = self.nhru
#print(self.nhru)
self.richards.m[:] = self.input_fp.groups['parameters'].variables['m'][:]
#self.richards.dem[:] = self.input_fp.groups['parameters'].variables['dem'][:]
self.richards.dem[:] = self.input_fp.groups['parameters'].variables['hand'][:]
self.richards.slope[:] = self.input_fp.groups['parameters'].variables['slope'][:]
#self.richards.hand[:] = self.input_fp.groups['parameters'].variables['hand'][:]
self.richards.area[:] = self.input_fp.groups['parameters'].variables['area'][:]
self.richards.width = sparse.csr_matrix((self.input_fp.groups['wmatrix'].variables['data'][:],
self.input_fp.groups['wmatrix'].variables['indices'][:],
self.input_fp.groups['wmatrix'].variables['indptr'][:]),
shape=(self.nhru,self.nhru),dtype=np.float64)
#dtype=np.float64)[0:self.nhru,0:self.nhru]
#print(self.richards.width.shape)
self.richards.I = self.richards.width.copy()
self.richards.I[self.richards.I != 0] = 1
return
def test_hydraulic_conductivity(self): nhru = 100 nsoil = 5 model = richards.richards(nhru,1) model.ksat[:] = 10**-3*np.random.uniform(low=10**-6,high=1,size=nhru) #mm/s model.satpsi[:] = np.random.uniform(low=0.01,high=1.0,size=nhru) model.b[:] = np.random.uniform(low=3,high=8,size=nhru) psi = np.random.uniform(low=0.001,high=10**10,size=nhru) #model kx = model.calculate_hydraulic_conductivity(psi) #baseline bkx = 10*model.ksat*(psi/model.satpsi)**(-2-3/model.b) self.assertTrue(np.allclose(kx,bkx,rtol=1e-10, atol=1e-50))
def test_hydraulic_conductivity(self):
nhru = 100
nsoil = 5
model = richards.richards(nhru, 1)
model.ksat[:] = 10**-3 * np.random.uniform(
low=10**-6, high=1, size=nhru) #mm/s
model.satpsi[:] = np.random.uniform(low=0.01, high=1.0, size=nhru)
model.b[:] = np.random.uniform(low=3, high=8, size=nhru)
psi = np.random.uniform(low=0.001, high=10**10, size=nhru)
#model
kx = model.calculate_hydraulic_conductivity(psi)
#baseline
bkx = 10 * model.ksat * (psi / model.satpsi)**(-2 - 3 / model.b)
self.assertTrue(np.allclose(kx, bkx, rtol=1e-10, atol=1e-50))
def test_dense_sparse_comparison(self):
nhru = 5000
nsoil = 1
model = richards.richards(nhru, nsoil)
#define parameters
model.nhru = nhru
model.theta[:] = np.random.uniform(low=0.1,
high=0.5,
size=(nhru, nsoil))
model.thetar[:] = 0.1
model.thetas[:] = 0.5
model.b[:] = np.random.uniform(low=0.1, high=0.5, size=nhru)
model.satpsi[:] = np.random.uniform(low=0.001, high=1.0, size=nhru)
model.ksat[:] = 10**-3 * np.random.uniform(
low=10**-6, high=1, size=nhru) #mm/s
model.dz[:] = np.random.uniform(low=0.01, high=100.0, size=nsoil)
model.dx = 30.0
model.area[:] = np.random.uniform(low=900, high=90000, size=nhru)
model.dem[:] = np.random.uniform(low=0.0, high=1000.0, size=nhru)
tmp = np.random.randint(low=-1000, high=2, size=(nhru, nhru))
tmp[tmp <= 0] = 0
tmp[range(nhru), range(nhru)] = 1
tsymm = 30 * (tmp + tmp.T) / 2
model.width = scipy.sparse.csr_matrix(tsymm)
model.I = model.width.copy()
model.I[model.I != 0] = 1
print '%d connections out of %d possible connections' % (np.sum(
model.width != 0), nhru * nhru)
#update (dense)
tic = time.time()
model.update(type='dense')
print 'dense', time.time() - tic
hdiv = np.copy(model.hdiv)
#update (sparse)
tic = time.time()
model.update(type='sparse')
print 'sparse', time.time() - tic
bhdiv = np.copy(model.hdiv)
#compare
self.assertTrue(np.allclose(hdiv, bhdiv, rtol=1e-10, atol=1e-50))
def test_dense_sparse_comparison(self): nhru = 5000 nsoil = 1 model = richards.richards(nhru,nsoil) #define parameters model.nhru = nhru model.theta[:] = np.random.uniform(low=0.1,high=0.5,size=(nhru,nsoil)) model.thetar[:] = 0.1 model.thetas[:] = 0.5 model.b[:] = np.random.uniform(low=0.1,high=0.5,size=nhru) model.satpsi[:] = np.random.uniform(low=0.001,high=1.0,size=nhru) model.ksat[:] = 10**-3*np.random.uniform(low=10**-6,high=1,size=nhru) #mm/s model.dz[:] = np.random.uniform(low=0.01,high=100.0,size=nsoil) model.dx = 30.0 model.area[:] = np.random.uniform(low=900,high=90000,size=nhru) model.dem[:] = np.random.uniform(low=0.0,high=1000.0,size=nhru) tmp = np.random.randint(low=-1000,high=2,size=(nhru,nhru)) tmp[tmp <= 0] = 0 tmp[range(nhru),range(nhru)] = 1 tsymm = 30*(tmp + tmp.T)/2 model.width = scipy.sparse.csr_matrix(tsymm) model.I = model.width.copy() model.I[model.I != 0] = 1 print '%d connections out of %d possible connections' % (np.sum(model.width != 0),nhru*nhru) #update (dense) tic = time.time() model.update(type='dense') print 'dense',time.time()-tic hdiv = np.copy(model.hdiv) #update (sparse) tic = time.time() model.update(type='sparse') print 'sparse',time.time()-tic bhdiv = np.copy(model.hdiv) #compare self.assertTrue(np.allclose(hdiv,bhdiv,rtol=1e-10, atol=1e-50))
def test_soil_moisture_potential(self):
eps = 0.01
nhru = 100
nsoil = 5
model = richards.richards(nhru,nsoil)
#define parameters
model.theta[:] = np.random.uniform(low=0.05,high=0.5,size=nhru)[:,np.newaxis]
model.thetar[:] = 0.1
model.thetas[:] = 0.5
model.b[:] = np.random.uniform(low=3,high=8,size=nhru)
model.satpsi[:] = 0.1
#choose layer
psi = []
bpsi = []
for il in xrange(nsoil):
psi.append(model.calculate_soil_moisture_potential(il))
tmp = model.theta[:,il]
m = (tmp <= (1+eps)*model.thetar)
tmp[m] = (1+eps)*model.thetar[m]
bpsi.append(model.satpsi*((model.theta[:,il]-model.thetar)/(model.thetas-model.thetar))**(-model.b))
psi = np.array(psi)
bpsi = np.array(bpsi)
self.assertTrue(np.allclose(psi,bpsi))