def test_degenerate_drainage(): """ This "hourglass" configuration should be one of the hardest to correctly re-route. """ mg = RasterModelGrid(9, 5) z_init = mg.node_x.copy()*0.0001 + 1. lake_pits = np.array([7, 11, 12, 13, 17, 27, 31, 32, 33, 37]) z_init[lake_pits] = -1. z_init[22] = 0. # the common spill pt for both lakes z_init[21] = 0.1 # an adverse bump in the spillway z_init[20] = -0.2 # the spillway z = mg.add_field('node', 'topographic__elevation', z_init) fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() lf.map_depressions() correct_A = np.array([ 0., 0., 0., 0., 0., 0., 1., 3., 1., 0., 0., 5., 1., 2., 0., 0., 1., 10., 1., 0., 21., 21., 1., 1., 0., 0., 1., 9., 1., 0., 0., 3., 1., 2., 0., 0., 1., 1., 1., 0., 0., 0., 0., 0., 0.]) thelake = np.concatenate((lake_pits, [22])).sort() assert_array_almost_equal(mg.at_node['drainage_area'], correct_A)
def test_three_pits(): """ A test to ensure the component correctly handles cases where there are multiple pits. """ mg = RasterModelGrid(10,10,1.) z = mg.add_field('node', 'topographic__elevation', mg.node_x.copy()) # a sloping plane #np.random.seed(seed=0) #z += np.random.rand(100)/10000. # punch some holes z[33] = 1. z[43] = 1. z[37] = 4. z[74:76] = 1. fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() lf.map_depressions() flow_sinks_target = np.zeros(100, dtype=bool) flow_sinks_target[mg.boundary_nodes] = True # no internal sinks now: assert_array_equal(mg.at_node['flow__sink_flag'], flow_sinks_target) # test conservation of mass: assert_almost_equal(mg.at_node['drainage_area' ].reshape((10,10))[1:-1,1].sum(), 8.**2) # ^all the core nodes # test the actual flow field: nA = np.array([ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 8., 8., 7., 6., 5., 4., 3., 2., 1., 0., 2., 2., 1., 1., 2., 1., 1., 1., 1., 0., 26., 26., 25., 15., 11., 10., 9., 8., 1., 0., 2., 2., 1., 9., 2., 1., 1., 1., 1., 0., 2., 2., 1., 1., 5., 4., 3., 2., 1., 0., 2., 2., 1., 1., 1., 1., 3., 2., 1., 0., 20., 20., 19., 18., 17., 12., 3., 2., 1., 0., 2., 2., 1., 1., 1., 1., 3., 2., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]) assert_array_equal(mg.at_node['drainage_area'], nA) #test a couple more properties: lc = np.empty(100, dtype=int) lc.fill(XX) lc[33] = 33 lc[43] = 33 lc[37] = 37 lc[74:76] = 74 assert_array_equal(lf.lake_map, lc) assert_array_equal(lf.lake_codes, [33, 37, 74]) assert_equal(lf.number_of_lakes, 3) assert_array_almost_equal(lf.lake_areas, [2., 1., 2.]) assert_array_almost_equal(lf.lake_volumes, [2., 2., 4.])
def test_lake_mapper(): """ Create a test grid and run a series of tests. """ # Make a test grid rmg = create_test_grid() # Instantiate a lake mapper # (Note that we don't need to send it an input file name, because our grid # already has a topographic__elevation field) lm = DepressionFinderAndRouter(rmg) # Run it on our test grid lm.map_depressions() # Run tests check_fields1(rmg) check_array_values1(rmg, lm) check_fields2(rmg) check_array_values2(rmg, lm)
def test_composite_pits(): """ A test to ensure the component correctly handles cases where there are multiple pits, inset into each other. """ mg = RasterModelGrid(10, 10, 1.) z = mg.add_field('node', 'topographic__elevation', mg.node_x.copy()) # a sloping plane #np.random.seed(seed=0) #z += np.random.rand(100)/10000. # punch one big hole z.reshape((10,10))[3:8,3:8] = 0. # dig a couple of inset holes z[57] = -1. z[44] = -2. z[54] = -10. fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() lf.map_depressions() flow_sinks_target = np.zeros(100, dtype=bool) flow_sinks_target[mg.boundary_nodes] = True # no internal sinks now: assert_array_equal(mg.at_node['flow__sink_flag'], flow_sinks_target) # test conservation of mass: assert_almost_equal(mg.at_node['drainage_area' ].reshape((10,10))[1:-1,1].sum(), 8.**2) # ^all the core nodes # test the actual flow field: nA = np.array([ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 8., 8., 7., 6., 5., 4., 3., 2., 1., 0., 1., 1., 1., 1., 1., 1., 1., 1., 1., 0., 1., 1., 1., 4., 2., 2., 8., 4., 1., 0., 1., 1., 1., 8., 3., 15., 3., 2., 1., 0., 1., 1., 1., 13., 25., 6., 3., 2., 1., 0., 1., 1., 1., 45., 3., 3., 5., 2., 1., 0., 50., 50., 49., 3., 2., 2., 2., 4., 1., 0., 1., 1., 1., 1., 1., 1., 1., 1., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]) assert_array_equal(mg.at_node['drainage_area'], nA) # the lake code map: lc = np.array([XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, 57, 57, 57, 57, 57, XX, XX, XX, XX, XX, 57, 57, 57, 57, 57, XX, XX, XX, XX, XX, 57, 57, 57, 57, 57, XX, XX, XX, XX, XX, 57, 57, 57, 57, 57, XX, XX, XX, XX, XX, 57, 57, 57, 57, 57, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX]) #test the remaining properties: assert_equal(lf.lake_outlets.size, 1) assert_equal(lf.lake_outlets[0], 72) outlets_in_map = np.unique(lf.depression_outlet_map) assert_equal(outlets_in_map.size, 2) assert_equal(outlets_in_map[1], 72) assert_equal(lf.number_of_lakes, 1) assert_equal(lf.lake_codes[0], 57) assert_array_equal(lf.lake_map, lc) assert_almost_equal(lf.lake_areas[0], 25.) assert_almost_equal(lf.lake_volumes[0], 63.)
d = 0.1 #z0 tol = 0.01 #m water thickness error allowable elapsed_time = 0. keep_running = True counter = 0 # simple incremented counter to let us see the model advance while keep_running: if elapsed_time + dt > runtime: dt = runtime - elapsed_time keep_running = False #route flow to determine water volume flux in each cell _ = fr.route_flow() # route_flow isn't time sensitive, so it doesn't take dt as input #lake filling stuff here try: _ = lf.map_depressions() #_ = lf.route_flow() #route flow given that lakes may exist except AssertionError: print 'Be careful, depression filler took too many iterations' print counter print np.amin(mg['node']['topographic__elevation']) print '-----' #test = mg['node']['upstream_ID_order'] #sys.exit('f**k you') #turn volume flux into specific discharge by dividing by channel width water_vol_flux = mg['node']['water__volume_flux'] #/ 365 / 24 / 3600 #drainage area times rain rate water_specific_q = water_vol_flux / channel_width fluvial_nodes = np.where(mg.at_node['drainage_area'] >= threshold_drainage_area)[0] for x in range(len(fluvial_nodes)): s = mg['node']['topographic__steepest_slope'][fluvial_nodes[x]]
np.vstack((mg.node_x.max() - mg.node_x, mg.node_x - mg.node_x.min())), axis=0 ) y_distance_from_edge = np.amin( np.vstack((mg.node_y.max() - mg.node_y, mg.node_y - mg.node_y.min())), axis=0 ) # make the "hole" hole_elev = np.sqrt(x_distance_from_center ** 2 + y_distance_from_center ** 2) # make the rim: rim_elev = np.sqrt(x_distance_from_edge ** 2 + y_distance_from_edge ** 2) # assemble z = np.amin(np.vstack((hole_elev, rim_elev)), axis=0) z += np.random.rand(nx * ny) / 1000. mg.add_field("node", "topographic__elevation", z, copy=False) fr = FlowAccumulator(mg, flow_director="D8") lf = DepressionFinderAndRouter(mg) fr.run_one_step() figure("old drainage area") imshow_grid(mg, "drainage_area") lf.map_depressions(pits=mg.at_node["flow__sink_flag"]) figure("depression depth") imshow_grid(mg, "depression__depth") figure("new drainage area") imshow_grid(mg, "drainage_area")
class SinkFiller(Component): """ This component identifies depressions in a topographic surface, then fills them in in the topography. No attempt is made to conserve sediment mass. User may specify whether the holes should be filled to flat, or with a gradient downwards towards the depression outlet. The gradient can be spatially variable, and is chosen to not reverse any drainage directions at the perimeter of each lake. The primary method of this class is 'run_one_step'. 'fill_pits' is a synonym. Constructor assigns a copy of the grid, and calls the initialize method. Construction:: SinkFiller(grid, routing='D8', apply_slope=False, fill_slope=1.e-5): Parameters ---------- grid : ModelGrid A landlab grid. routing : {'D8', 'D4'} (optional) If grid is a raster type, controls whether fill connectivity can occur on diagonals ('D8', default), or only orthogonally ('D4'). Has no effect if grid is not a raster. apply_slope : bool If False (default), leave the top of the filled sink flat. If True, apply the slope fill_slope to the top surface to allow subsequent flow routing. A test is performed to ensure applying this slope will not alter the drainage structure at the edge of the filled region (i.e., that we are not accidentally reversing the flow direction far from the outlet.) fill_slope : float (m/m) The slope added to the top surface of filled pits to allow flow routing across them, if apply_slope. Examples -------- >>> from landlab import RasterModelGrid >>> from landlab import BAD_INDEX_VALUE as XX >>> from landlab.components import FlowRouter, SinkFiller >>> import numpy as np >>> lake1 = np.array([34, 35, 36, 44, 45, 46, 54, 55, 56, 65, 74]) >>> lake2 = np.array([78, 87, 88]) >>> guard_nodes = np.array([23, 33, 53, 63, 73, 83]) >>> lake = np.concatenate((lake1, lake2)) >>> mg = RasterModelGrid((10, 10), 1.) >>> z = np.ones(100, dtype=float) >>> z += mg.node_x # add a slope >>> z[guard_nodes] += 0.001 # forces the flow out of a particular node >>> z[lake] = 0. >>> field = mg.add_field('node', 'topographic__elevation', z, ... units='-', copy=True) >>> fr = FlowRouter(mg) >>> fr.run_one_step() >>> mg.at_node['flow__sink_flag'][mg.core_nodes].sum() 14 >>> hf = SinkFiller(mg, apply_slope=False) >>> hf.run_one_step() >>> np.allclose(mg.at_node['topographic__elevation'][lake1], 4.) True >>> np.allclose(mg.at_node['topographic__elevation'][lake2], 7.) True Now reset and demonstrate the adding of an inclined surface: >>> field[:] = z >>> hf = SinkFiller(mg, apply_slope=True) >>> hf.run_one_step() >>> hole1 = np.array([4.00007692, 4.00015385, 4.00023077, 4.00053846, ... 4.00038462, 4.00030769, 4.00069231, 4.00061538, ... 4.00046154, 4.00076923, 4.00084615]) >>> hole2 = np.array([7.4, 7.2, 7.6]) >>> np.allclose(mg.at_node['topographic__elevation'][lake1], hole1) True >>> np.allclose(mg.at_node['topographic__elevation'][lake2], hole2) True >>> fr.run_one_step() >>> mg.at_node['flow__sink_flag'][mg.core_nodes].sum() 0 """ _name = 'SinkFiller' _input_var_names = ('topographic__elevation', ) _output_var_names = ( 'topographic__elevation', 'sediment_fill__depth', ) _var_units = { 'topographic__elevation': 'm', 'sediment_fill__depth': 'm', } _var_mapping = { 'topographic__elevation': 'node', 'sediment_fill__depth': 'node', } _var_doc = { 'topographic__elevation': 'Surface topographic elevation', 'sediment_fill__depth': 'Depth of sediment added at each' + 'node', } @use_file_name_or_kwds def __init__(self, grid, routing='D8', apply_slope=False, fill_slope=1.e-5, **kwds): self._grid = grid if routing is not 'D8': assert routing is 'D4' self._routing = routing if ((type(self._grid) is landlab.grid.raster.RasterModelGrid) and (routing is 'D8')): self._D8 = True self.num_nbrs = 8 else: self._D8 = False # useful shorthand for thia test we do a lot if type(self._grid) is landlab.grid.raster.RasterModelGrid: self.num_nbrs = 4 self._fill_slope = fill_slope self._apply_slope = apply_slope self.initialize() def initialize(self, input_stream=None): """ The BMI-style initialize method takes an optional input_stream parameter, which may be either a ModelParameterDictionary object or an input stream from which a ModelParameterDictionary can read values. """ # Create a ModelParameterDictionary for the inputs if input_stream is None: inputs = None elif type(input_stream) == ModelParameterDictionary: inputs = input_stream else: inputs = ModelParameterDictionary(input_stream) # Make sure the grid includes elevation data. This means either: # 1. The grid has a node field called 'topographic__elevation', or # 2. The input file has an item called 'ELEVATION_FIELD_NAME' *and* # a field by this name exists in the grid. try: self._elev = self._grid.at_node['topographic__elevation'] except FieldError: try: self.topo_field_name = inputs.read_string('ELEVATION_' + 'FIELD_NAME') except AttributeError: print('Error: Because your grid does not have a node field') print('called "topographic__elevation", you need to pass the') print('name of a text input file or ModelParameterDictionary,') print('and this file or dictionary needs to include the name') print('of another field in your grid that contains your') print('elevation data.') raise AttributeError except MissingKeyError: print('Error: Because your grid does not have a node field') print('called "topographic__elevation", your input file (or') print('ModelParameterDictionary) must include an entry with') print('the key "ELEVATION_FIELD_NAME", which gives the name') print('of a field in your grid that contains your elevation') print('data.') raise MissingKeyError('ELEVATION_FIELD_NAME') try: self._elev = self._grid.at_node[self.topo_field_name] except AttributeError: print('Your grid does not seem to have a node field called', self.topo_field_name) else: self.topo_field_name = 'topographic__elevation' # create the only new output field: self.sed_fill_depth = self._grid.add_zeros('node', 'sediment_fill__depth', noclobber=False) self._lf = DepressionFinderAndRouter(self._grid, routing=self._routing) self._fr = FlowRouter(self._grid, method=self._routing) def fill_pits(self, **kwds): """ This is a synonym for the main method :func:`run_one_step`. """ self.run_one_step(**kwds) def run_one_step(self, **kwds): """ This is the main method. Call it to fill depressions in a starting topography. """ # added for back-compatibility with old formats try: self._apply_slope = kwds['apply_slope'] except KeyError: pass self.original_elev = self._elev.copy() # We need this, as we'll have to do ALL this again if we manage # to jack the elevs too high in one of the "subsidiary" lakes. # We're going to implement the lake_mapper component to do the heavy # lifting here, then delete its fields. This means we first need to # test if these fields already exist, in which case, we should *not* # delete them! existing_fields = {} spurious_fields = set() set_of_outputs = (set(self._lf.output_var_names) | set(self._fr.output_var_names)) try: set_of_outputs.remove(self.topo_field_name) except KeyError: pass for field in set_of_outputs: try: existing_fields[field] = self._grid.at_node[field].copy() except FieldError: # not there; good! spurious_fields.add(field) self._fr.route_flow() self._lf.map_depressions(pits=self._grid.at_node['flow__sink_flag'], reroute_flow=True) # add the depression depths to get up to flat: self._elev += self._grid.at_node['depression__depth'] # if apply_slope is none, we're now done! But if not... if self._apply_slope: # new way of doing this - use the upstream structure! Should be # both more general and more efficient for (outlet_node, lake_code) in zip(self._lf.lake_outlets, self._lf.lake_codes): lake_nodes = np.where(self._lf.lake_map == lake_code)[0] lake_perim = self._get_lake_ext_margin(lake_nodes) perim_elevs = self._elev[lake_perim] out_elev = self._elev[outlet_node] lowest_elev_perim = perim_elevs[perim_elevs != out_elev].min() # note we exclude the outlet node elev_increment = ( (lowest_elev_perim - self._elev[outlet_node]) / (lake_nodes.size + 2.)) assert elev_increment > 0. all_ordering = self._grid.at_node['flow__upstream_node_order'] upstream_order_bool = np.in1d(all_ordering, lake_nodes, assume_unique=True) lake_upstream_order = all_ordering[upstream_order_bool] argsort_lake = np.argsort(lake_upstream_order) elevs_to_add = (np.arange(lake_nodes.size, dtype=float) + 1.) * elev_increment sorted_elevs_to_add = elevs_to_add[argsort_lake] self._elev[lake_nodes] += sorted_elevs_to_add # now put back any fields that were present initially, and wipe the # rest: for delete_me in spurious_fields: self._grid.delete_field('node', delete_me) for update_me in existing_fields.keys(): self.grid.at_node[update_me][:] = existing_fields[update_me] # fill the output field self.sed_fill_depth[:] = self._elev - self.original_elev @deprecated(use='fill_pits', version=1.0) def _fill_pits_old(self, apply_slope=None): """ .. deprecated:: 0.1.38 Use :func:`fill_pits` instead. This is the main method. Call it to fill depressions in a starting topography. **Output fields** * `topographic__elevation` : the updated elevations * `sediment_fill__depth` : the depth of sediment added at each node Parameters ---------- apply_slope : None, bool, or float If a float is provided this is the slope of the surface down towards the lake outlet. Supply a small positive number, e.g., 1.e-5 (or True, to use this default value). A test is performed to ensure applying this slope will not alter the drainage structure at the edge of the filled region (i.e., that we are not accidentally reversing the flow direction far from the outlet.) The component will automatically decrease the (supplied or default) gradient a number of times to try to accommodate this, but will eventually raise an OverflowError if it can't deal with it. If you pass True, the method will use the default value of 1.e-5. """ self.original_elev = self._elev.copy() # We need this, as we'll have to do ALL this again if we manage # to jack the elevs too high in one of the "subsidiary" lakes. # We're going to implement the lake_mapper component to do the heavy # lifting here, then delete its fields. This means we first need to # test if these fields already exist, in which case, we should *not* # delete them! existing_fields = {} spurious_fields = set() set_of_outputs = self._lf.output_var_names | self._fr.output_var_names try: set_of_outputs.remove(self.topo_field_name) except KeyError: pass for field in set_of_outputs: try: existing_fields[field] = self._grid.at_node[field].copy() except FieldError: # not there; good! spurious_fields.add(field) self._fr.route_flow() self._lf.map_depressions(pits=self._grid.at_node['flow__sink_flag'], reroute_flow=False) # add the depression depths to get up to flat: self._elev += self._grid.at_node['depression__depth'] # if apply_slope is none, we're now done! But if not... if apply_slope is True: apply_slope = self._fill_slope elif type(apply_slope) in (float, int): assert apply_slope >= 0. if apply_slope: # this isn't very efficient, but OK as we're only running this # code ONCE in almost all use cases sublake = False unstable = True stability_increment = 0 self.lake_nodes_treated = np.array([], dtype=int) while unstable: while 1: for (outlet_node, lake_code) in zip(self._lf.lake_outlets, self._lf.lake_codes): self._apply_slope_current_lake(apply_slope, outlet_node, lake_code, sublake) # Call the mapper again here. Bail out if no core pits are # found. # This is necessary as there are some configs where adding # the slope could create subsidiary pits in the topo self._lf.map_depressions(pits=None, reroute_flow=False) if len(self._lf.lake_outlets) == 0.: break self._elev += self._grid.at_node['depression__depth'] sublake = True self.lake_nodes_treated = np.array([], dtype=int) # final test that all lakes are not reversing flow dirs all_lakes = np.where( self._lf.flood_status < BAD_INDEX_VALUE)[0] unstable = self.drainage_directions_change( all_lakes, self.original_elev, self._elev) if unstable: apply_slope *= 0.1 sublake = False self.lake_nodes_treated = np.array([], dtype=int) self._elev[:] = original_elev # put back init conds stability_increment += 1 if stability_increment == 10: raise OverflowError('Filler could not find a stable ' + 'condition with a sloping ' + 'surface!') # now put back any fields that were present initially, and wipe the # rest: for delete_me in spurious_fields: self._grid.delete_field('node', delete_me) for update_me in existing_fields.keys(): self.grid.at_node[update_me] = existing_fields[update_me] # fill the output field self.sed_fill_depth[:] = self._elev - self.original_elev def _add_slopes(self, slope, outlet_node, lake_code): """ Assuming you have already run the lake_mapper, adds an incline towards the outlet to the nodes in the lake. """ new_elevs = self._elev.copy() outlet_coord = (self._grid.node_x[outlet_node], self._grid.node_y[outlet_node]) lake_nodes = np.where(self._lf.lake_map == lake_code)[0] lake_nodes = np.setdiff1d(lake_nodes, self.lake_nodes_treated) lake_ext_margin = self._get_lake_ext_margin(lake_nodes) d = self._grid.calc_distances_of_nodes_to_point(outlet_coord, node_subset=lake_nodes) add_vals = slope * d new_elevs[lake_nodes] += add_vals self.lake_nodes_treated = np.union1d(self.lake_nodes_treated, lake_nodes) return new_elevs, lake_nodes def _get_lake_ext_margin(self, lake_nodes): """ Returns the nodes forming the external margin of the lake, honoring the *routing* method (D4/D8) if applicable. """ if self._D8 is True: all_poss = np.union1d( self._grid.active_neighbors_at_node(lake_nodes), self._grid._get_diagonal_list(lake_nodes)) else: all_poss = np.unique( self._grid.active_neighbors_at_node(lake_nodes)) lake_ext_edge = np.setdiff1d(all_poss, lake_nodes) return lake_ext_edge[lake_ext_edge != BAD_INDEX_VALUE] def _get_lake_int_margin(self, lake_nodes, lake_ext_edge): """ Returns the nodes forming the internal margin of the lake, honoring the *routing* method (D4/D8) if applicable. """ lee = lake_ext_edge if self._D8 is True: all_poss_int = np.union1d(self._grid.active_neighbors_at_node(lee), self._grid._get_diagonal_list(lee)) else: all_poss_int = np.unique(self._grid.active_neighbors_at_node(lee)) lake_int_edge = np.intersect1d(all_poss_int, lake_nodes) return lake_int_edge[lake_int_edge != BAD_INDEX_VALUE] def _apply_slope_current_lake(self, apply_slope, outlet_node, lake_code, sublake): """ Wraps the _add_slopes method to allow handling of conditions where the drainage structure would be changed or we're dealing with a sublake. """ while 1: starting_elevs = self._elev.copy() self._elev[:], lake_nodes = self._add_slopes( apply_slope, outlet_node, lake_code) ext_edge = self._get_lake_ext_margin(lake_nodes) if sublake: break else: if not self.drainage_directions_change( lake_nodes, starting_elevs, self._elev): break else: # put the elevs back... self._elev[lake_nodes] = starting_elevs[lake_nodes] # the slope was too big. Reduce it. apply_slope *= 0.1 # if we get here, either sublake, or drainage dirs are stable def drainage_directions_change(self, lake_nodes, old_elevs, new_elevs): """ True if the drainage structure at lake margin changes, False otherwise. """ ext_edge = self._get_lake_ext_margin(lake_nodes) if self._D8: edge_neighbors = np.hstack( (self._grid.active_neighbors_at_node(ext_edge), self._grid._get_diagonal_list(ext_edge))) else: edge_neighbors = self._grid.active_neighbors_at_node( ext_edge).copy() edge_neighbors[edge_neighbors == BAD_INDEX_VALUE] = -1 # ^value irrelevant old_neighbor_elevs = old_elevs[edge_neighbors] new_neighbor_elevs = new_elevs[edge_neighbors] # enforce the "don't change drainage direction" condition: edge_elevs = old_elevs[ext_edge].reshape((ext_edge.size, 1)) cond = np.allclose((edge_elevs >= old_neighbor_elevs), (edge_elevs >= new_neighbor_elevs)) # if True, we're good, the tilting didn't mess with the fr return not cond
fr = FlowAccumulator(mg, flow_director="D8") sp = StreamPowerEroder(mg, "./drive_sp_params.txt") lf = DepressionFinderAndRouter(mg) # load the Fastscape module too, to allow direct comparison fsp = FastscapeEroder(mg, "./drive_sp_params.txt") # perform the loop: elapsed_time = 0. # total time in simulation while elapsed_time < time_to_run: # for i in range(10): print(elapsed_time) if elapsed_time + dt > time_to_run: print("Short step!") dt = time_to_run - elapsed_time mg = fr.route_flow(method="D8") lf.map_depressions() # print 'Area: ', numpy.max(mg.at_node['drainage_area']) # mg = fsp.erode(mg) mg, _, _ = sp.erode( mg, dt, node_drainage_areas="drainage_area", slopes_at_nodes="topographic__steepest_slope", K_if_used="K_values", ) # add uplift mg.at_node["topographic__elevation"][mg.core_nodes] += uplift * dt elapsed_time += dt time_off = time.time() print("Elapsed time: ", time_off - time_on)
(mg.node_x.max() - mg.node_x, mg.node_x - mg.node_x.min())), axis=0) y_distance_from_edge = np.amin(np.vstack( (mg.node_y.max() - mg.node_y, mg.node_y - mg.node_y.min())), axis=0) # make the "hole" hole_elev = np.sqrt(x_distance_from_center**2 + y_distance_from_center**2) # make the rim: rim_elev = np.sqrt(x_distance_from_edge**2 + y_distance_from_edge**2) # assemble z = np.amin(np.vstack((hole_elev, rim_elev)), axis=0) z += np.random.rand(nx * ny) / 1000. mg.add_field("node", "topographic__elevation", z, copy=False) fr = FlowAccumulator(mg, flow_director="D8") lf = DepressionFinderAndRouter(mg) fr.run_one_step() figure("old drainage area") imshow_node_grid(mg, "drainage_area") lf.map_depressions(pits=mg.at_node["flow__sink_flag"]) figure("depression depth") imshow_node_grid(mg, "depression__depth") figure("new drainage area") imshow_node_grid(mg, "drainage_area")
class SinkFiller(Component): """ This component identifies depressions in a topographic surface, then fills them in in the topography. No attempt is made to conserve sediment mass. User may specify whether the holes should be filled to flat, or with a gradient downwards towards the depression outlet. The gradient can be spatially variable, and is chosen to not reverse any drainage directions at the perimeter of each lake. The primary method of this class is 'run_one_step'. 'fill_pits' is a synonym. Constructor assigns a copy of the grid, and calls the initialize method. Construction:: SinkFiller(grid, routing='D8', apply_slope=False, fill_slope=1.e-5): Parameters ---------- grid : ModelGrid A landlab grid. routing : {'D8', 'D4'} (optional) If grid is a raster type, controls whether fill connectivity can occur on diagonals ('D8', default), or only orthogonally ('D4'). Has no effect if grid is not a raster. apply_slope : bool If False (default), leave the top of the filled sink flat. If True, apply the slope fill_slope to the top surface to allow subsequent flow routing. A test is performed to ensure applying this slope will not alter the drainage structure at the edge of the filled region (i.e., that we are not accidentally reversing the flow direction far from the outlet.) fill_slope : float (m/m) The slope added to the top surface of filled pits to allow flow routing across them, if apply_slope. Examples -------- >>> from landlab import RasterModelGrid >>> from landlab import BAD_INDEX_VALUE as XX >>> from landlab.components import FlowRouter, SinkFiller >>> import numpy as np >>> lake1 = np.array([34, 35, 36, 44, 45, 46, 54, 55, 56, 65, 74]) >>> lake2 = np.array([78, 87, 88]) >>> guard_nodes = np.array([23, 33, 53, 63, 73, 83]) >>> lake = np.concatenate((lake1, lake2)) >>> mg = RasterModelGrid((10, 10), 1.) >>> z = np.ones(100, dtype=float) >>> z += mg.node_x # add a slope >>> z[guard_nodes] += 0.001 # forces the flow out of a particular node >>> z[lake] = 0. >>> field = mg.add_field('node', 'topographic__elevation', z, ... units='-', copy=True) >>> fr = FlowRouter(mg) >>> fr.run_one_step() >>> mg.at_node['flow__sink_flag'][mg.core_nodes].sum() 14 >>> hf = SinkFiller(mg, apply_slope=False) >>> hf.run_one_step() >>> np.allclose(mg.at_node['topographic__elevation'][lake1], 4.) True >>> np.allclose(mg.at_node['topographic__elevation'][lake2], 7.) True Now reset and demonstrate the adding of an inclined surface: >>> field[:] = z >>> hf = SinkFiller(mg, apply_slope=True) >>> hf.run_one_step() >>> hole1 = np.array([4.00007692, 4.00015385, 4.00023077, 4.00053846, ... 4.00038462, 4.00030769, 4.00069231, 4.00061538, ... 4.00046154, 4.00076923, 4.00084615]) >>> hole2 = np.array([7.4, 7.2, 7.6]) >>> np.allclose(mg.at_node['topographic__elevation'][lake1], hole1) True >>> np.allclose(mg.at_node['topographic__elevation'][lake2], hole2) True >>> fr.run_one_step() >>> mg.at_node['flow__sink_flag'][mg.core_nodes].sum() 0 """ _name = 'SinkFiller' _input_var_names = ('topographic__elevation', ) _output_var_names = ('topographic__elevation', 'sediment_fill__depth', ) _var_units = {'topographic__elevation': 'm', 'sediment_fill__depth': 'm', } _var_mapping = {'topographic__elevation': 'node', 'sediment_fill__depth': 'node', } _var_doc = {'topographic__elevation': 'Surface topographic elevation', 'sediment_fill__depth': 'Depth of sediment added at each' + 'node', } @use_file_name_or_kwds def __init__(self, grid, routing='D8', apply_slope=False, fill_slope=1.e-5, **kwds): self._grid = grid if routing is not 'D8': assert routing is 'D4' self._routing = routing if ((type(self._grid) is landlab.grid.raster.RasterModelGrid) and (routing is 'D8')): self._D8 = True self.num_nbrs = 8 else: self._D8 = False # useful shorthand for thia test we do a lot if type(self._grid) is landlab.grid.raster.RasterModelGrid: self.num_nbrs = 4 self._fill_slope = fill_slope self._apply_slope = apply_slope self.initialize() def initialize(self, input_stream=None): """ The BMI-style initialize method takes an optional input_stream parameter, which may be either a ModelParameterDictionary object or an input stream from which a ModelParameterDictionary can read values. """ # Create a ModelParameterDictionary for the inputs if input_stream is None: inputs = None elif type(input_stream) == ModelParameterDictionary: inputs = input_stream else: inputs = ModelParameterDictionary(input_stream) # Make sure the grid includes elevation data. This means either: # 1. The grid has a node field called 'topographic__elevation', or # 2. The input file has an item called 'ELEVATION_FIELD_NAME' *and* # a field by this name exists in the grid. try: self._elev = self._grid.at_node['topographic__elevation'] except FieldError: try: self.topo_field_name = inputs.read_string('ELEVATION_' + 'FIELD_NAME') except AttributeError: print('Error: Because your grid does not have a node field') print('called "topographic__elevation", you need to pass the') print('name of a text input file or ModelParameterDictionary,') print('and this file or dictionary needs to include the name') print('of another field in your grid that contains your') print('elevation data.') raise AttributeError except MissingKeyError: print('Error: Because your grid does not have a node field') print('called "topographic__elevation", your input file (or') print('ModelParameterDictionary) must include an entry with') print('the key "ELEVATION_FIELD_NAME", which gives the name') print('of a field in your grid that contains your elevation') print('data.') raise MissingKeyError('ELEVATION_FIELD_NAME') try: self._elev = self._grid.at_node[self.topo_field_name] except AttributeError: print('Your grid does not seem to have a node field called', self.topo_field_name) else: self.topo_field_name = 'topographic__elevation' # create the only new output field: self.sed_fill_depth = self._grid.add_zeros('node', 'sediment_fill__depth', noclobber=False) self._lf = DepressionFinderAndRouter(self._grid, routing=self._routing) self._fr = FlowRouter(self._grid, method=self._routing) def fill_pits(self, **kwds): """ This is a synonym for the main method :func:`run_one_step`. """ self.run_one_step(**kwds) def run_one_step(self, **kwds): """ This is the main method. Call it to fill depressions in a starting topography. """ # added for back-compatibility with old formats try: self._apply_slope = kwds['apply_slope'] except KeyError: pass self.original_elev = self._elev.copy() # We need this, as we'll have to do ALL this again if we manage # to jack the elevs too high in one of the "subsidiary" lakes. # We're going to implement the lake_mapper component to do the heavy # lifting here, then delete its fields. This means we first need to # test if these fields already exist, in which case, we should *not* # delete them! existing_fields = {} spurious_fields = set() set_of_outputs = (set(self._lf.output_var_names) | set(self._fr.output_var_names)) try: set_of_outputs.remove(self.topo_field_name) except KeyError: pass for field in set_of_outputs: try: existing_fields[field] = self._grid.at_node[field].copy() except FieldError: # not there; good! spurious_fields.add(field) self._fr.route_flow() self._lf.map_depressions(pits=self._grid.at_node['flow__sink_flag'], reroute_flow=True) # add the depression depths to get up to flat: self._elev += self._grid.at_node['depression__depth'] # if apply_slope is none, we're now done! But if not... if self._apply_slope: # new way of doing this - use the upstream structure! Should be # both more general and more efficient for (outlet_node, lake_code) in zip(self._lf.lake_outlets, self._lf.lake_codes): lake_nodes = np.where(self._lf.lake_map == lake_code)[0] lake_perim = self._get_lake_ext_margin(lake_nodes) perim_elevs = self._elev[lake_perim] out_elev = self._elev[outlet_node] lowest_elev_perim = perim_elevs[perim_elevs != out_elev].min() # note we exclude the outlet node elev_increment = ((lowest_elev_perim-self._elev[outlet_node]) / (lake_nodes.size + 2.)) assert elev_increment > 0. all_ordering = self._grid.at_node['flow__upstream_node_order'] upstream_order_bool = np.in1d(all_ordering, lake_nodes, assume_unique=True) lake_upstream_order = all_ordering[upstream_order_bool] argsort_lake = np.argsort(lake_upstream_order) elevs_to_add = (np.arange(lake_nodes.size, dtype=float) + 1.) * elev_increment sorted_elevs_to_add = elevs_to_add[argsort_lake] self._elev[lake_nodes] += sorted_elevs_to_add # now put back any fields that were present initially, and wipe the # rest: for delete_me in spurious_fields: self._grid.delete_field('node', delete_me) for update_me in existing_fields.keys(): self.grid.at_node[update_me][:] = existing_fields[update_me] # fill the output field self.sed_fill_depth[:] = self._elev - self.original_elev @deprecated(use='fill_pits', version=1.0) def _fill_pits_old(self, apply_slope=None): """ .. deprecated:: 0.1.38 Use :func:`fill_pits` instead. This is the main method. Call it to fill depressions in a starting topography. **Output fields** * `topographic__elevation` : the updated elevations * `sediment_fill__depth` : the depth of sediment added at each node Parameters ---------- apply_slope : None, bool, or float If a float is provided this is the slope of the surface down towards the lake outlet. Supply a small positive number, e.g., 1.e-5 (or True, to use this default value). A test is performed to ensure applying this slope will not alter the drainage structure at the edge of the filled region (i.e., that we are not accidentally reversing the flow direction far from the outlet.) The component will automatically decrease the (supplied or default) gradient a number of times to try to accommodate this, but will eventually raise an OverflowError if it can't deal with it. If you pass True, the method will use the default value of 1.e-5. """ self.original_elev = self._elev.copy() # We need this, as we'll have to do ALL this again if we manage # to jack the elevs too high in one of the "subsidiary" lakes. # We're going to implement the lake_mapper component to do the heavy # lifting here, then delete its fields. This means we first need to # test if these fields already exist, in which case, we should *not* # delete them! existing_fields = {} spurious_fields = set() set_of_outputs = self._lf.output_var_names | self._fr.output_var_names try: set_of_outputs.remove(self.topo_field_name) except KeyError: pass for field in set_of_outputs: try: existing_fields[field] = self._grid.at_node[field].copy() except FieldError: # not there; good! spurious_fields.add(field) self._fr.route_flow() self._lf.map_depressions(pits=self._grid.at_node['flow__sink_flag'], reroute_flow=False) # add the depression depths to get up to flat: self._elev += self._grid.at_node['depression__depth'] # if apply_slope is none, we're now done! But if not... if apply_slope is True: apply_slope = self._fill_slope elif type(apply_slope) in (float, int): assert apply_slope >= 0. if apply_slope: # this isn't very efficient, but OK as we're only running this # code ONCE in almost all use cases sublake = False unstable = True stability_increment = 0 self.lake_nodes_treated = np.array([], dtype=int) while unstable: while 1: for (outlet_node, lake_code) in zip(self._lf.lake_outlets, self._lf.lake_codes): self._apply_slope_current_lake(apply_slope, outlet_node, lake_code, sublake) # Call the mapper again here. Bail out if no core pits are # found. # This is necessary as there are some configs where adding # the slope could create subsidiary pits in the topo self._lf.map_depressions(pits=None, reroute_flow=False) if len(self._lf.lake_outlets) == 0.: break self._elev += self._grid.at_node['depression__depth'] sublake = True self.lake_nodes_treated = np.array([], dtype=int) # final test that all lakes are not reversing flow dirs all_lakes = np.where(self._lf.flood_status < BAD_INDEX_VALUE)[0] unstable = self.drainage_directions_change(all_lakes, self.original_elev, self._elev) if unstable: apply_slope *= 0.1 sublake = False self.lake_nodes_treated = np.array([], dtype=int) self._elev[:] = original_elev # put back init conds stability_increment += 1 if stability_increment == 10: raise OverflowError('Filler could not find a stable ' + 'condition with a sloping ' + 'surface!') # now put back any fields that were present initially, and wipe the # rest: for delete_me in spurious_fields: self._grid.delete_field('node', delete_me) for update_me in existing_fields.keys(): self.grid.at_node[update_me] = existing_fields[update_me] # fill the output field self.sed_fill_depth[:] = self._elev - self.original_elev def _add_slopes(self, slope, outlet_node, lake_code): """ Assuming you have already run the lake_mapper, adds an incline towards the outlet to the nodes in the lake. """ new_elevs = self._elev.copy() outlet_coord = (self._grid.node_x[outlet_node], self._grid.node_y[outlet_node]) lake_nodes = np.where(self._lf.lake_map == lake_code)[0] lake_nodes = np.setdiff1d(lake_nodes, self.lake_nodes_treated) lake_ext_margin = self._get_lake_ext_margin(lake_nodes) d = self._grid.calc_distances_of_nodes_to_point(outlet_coord, node_subset=lake_nodes) add_vals = slope*d new_elevs[lake_nodes] += add_vals self.lake_nodes_treated = np.union1d(self.lake_nodes_treated, lake_nodes) return new_elevs, lake_nodes def _get_lake_ext_margin(self, lake_nodes): """ Returns the nodes forming the external margin of the lake, honoring the *routing* method (D4/D8) if applicable. """ if self._D8 is True: all_poss = np.union1d(self._grid.active_neighbors_at_node(lake_nodes), self._grid._get_diagonal_list(lake_nodes)) else: all_poss = np.unique(self._grid.active_neighbors_at_node( lake_nodes)) lake_ext_edge = np.setdiff1d(all_poss, lake_nodes) return lake_ext_edge[lake_ext_edge != BAD_INDEX_VALUE] def _get_lake_int_margin(self, lake_nodes, lake_ext_edge): """ Returns the nodes forming the internal margin of the lake, honoring the *routing* method (D4/D8) if applicable. """ lee = lake_ext_edge if self._D8 is True: all_poss_int = np.union1d(self._grid.active_neighbors_at_node(lee), self._grid._get_diagonal_list(lee)) else: all_poss_int = np.unique(self._grid.active_neighbors_at_node(lee)) lake_int_edge = np.intersect1d(all_poss_int, lake_nodes) return lake_int_edge[lake_int_edge != BAD_INDEX_VALUE] def _apply_slope_current_lake(self, apply_slope, outlet_node, lake_code, sublake): """ Wraps the _add_slopes method to allow handling of conditions where the drainage structure would be changed or we're dealing with a sublake. """ while 1: starting_elevs = self._elev.copy() self._elev[:], lake_nodes = self._add_slopes(apply_slope, outlet_node, lake_code) ext_edge = self._get_lake_ext_margin(lake_nodes) if sublake: break else: if not self.drainage_directions_change(lake_nodes, starting_elevs, self._elev): break else: # put the elevs back... self._elev[lake_nodes] = starting_elevs[lake_nodes] # the slope was too big. Reduce it. apply_slope *= 0.1 # if we get here, either sublake, or drainage dirs are stable def drainage_directions_change(self, lake_nodes, old_elevs, new_elevs): """ True if the drainage structure at lake margin changes, False otherwise. """ ext_edge = self._get_lake_ext_margin(lake_nodes) if self._D8: edge_neighbors = np.hstack((self._grid.active_neighbors_at_node(ext_edge), self._grid._get_diagonal_list( ext_edge))) else: edge_neighbors = self._grid.active_neighbors_at_node(ext_edge).copy() edge_neighbors[edge_neighbors == BAD_INDEX_VALUE] = -1 # ^value irrelevant old_neighbor_elevs = old_elevs[edge_neighbors] new_neighbor_elevs = new_elevs[edge_neighbors] # enforce the "don't change drainage direction" condition: edge_elevs = old_elevs[ext_edge].reshape((ext_edge.size, 1)) cond = np.allclose((edge_elevs >= old_neighbor_elevs), (edge_elevs >= new_neighbor_elevs)) # if True, we're good, the tilting didn't mess with the fr return not cond
y_distance_from_center = mg.node_y-mg.node_y.mean() x_distance_from_edge = np.amin(np.vstack((mg.node_x.max()-mg.node_x, mg.node_x-mg.node_x.min())), axis=0) y_distance_from_edge = np.amin(np.vstack((mg.node_y.max()-mg.node_y, mg.node_y-mg.node_y.min())), axis=0) # make the "hole" hole_elev = np.sqrt(x_distance_from_center**2 + y_distance_from_center**2) # make the rim: rim_elev = np.sqrt(x_distance_from_edge**2 + y_distance_from_edge**2) # assemble z = np.amin(np.vstack((hole_elev, rim_elev)), axis=0) z += np.random.rand(nx*ny)/1000. mg.add_field('node', 'topographic__elevation', z, copy=False) fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() figure('old drainage area') imshow_node_grid(mg, 'drainage_area') lf.map_depressions(pits=mg.at_node['flow_sinks']) figure('depression depth') imshow_node_grid(mg, 'depression__depth') figure('new drainage area') imshow_node_grid(mg, 'drainage_area')
def test_edge_draining(): """ This tests when the lake attempts to drain from an edge, where an issue is suspected. """ # Create a 7x7 test grid with a well defined hole in it, AT THE EDGE. mg = RasterModelGrid((7, 7), (1., 1.)) z = mg.node_x.copy() guard_sides = np.concatenate((np.arange(7, 14), np.arange(35, 42))) edges = np.concatenate((np.arange(7), np.arange(42, 49))) hole_here = np.array(([15, 16, 22, 23, 29, 30])) z[guard_sides] = z[13] z[edges] = -2. # force flow outwards from the tops of the guards z[hole_here] = -1. A_new = np.array( [ [ [ 0., 1., 1., 1., 1., 1., 0., 0., 1., 1., 1., 1., 1., 0., 15., 5., 4., 3., 2., 1., 0., 0., 10., 4., 3., 2., 1., 0., 0., 1., 4., 3., 2., 1., 0., 0., 1., 1., 1., 1., 1., 0., 0., 1., 1., 1., 1., 1., 0., ] ] ] ).flatten() depr_outlet_target = np.array( [ XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, ] ).flatten() mg.add_field("node", "topographic__elevation", z, units="-") fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() lf.map_depressions() assert mg.at_node["drainage_area"] == approx(A_new) assert lf.depression_outlet_map == approx(depr_outlet_target)
def test_edge_draining(): """ This tests when the lake attempts to drain from an edge, where an issue is suspected. """ # Create a 7x7 test grid with a well defined hole in it, AT THE EDGE. mg = RasterModelGrid((7, 7), (1., 1.)) z = mg.node_x.copy() guard_sides = np.concatenate((np.arange(7, 14), np.arange(35, 42))) edges = np.concatenate((np.arange(7), np.arange(42, 49))) hole_here = np.array(([15, 16, 22, 23, 29, 30])) z[guard_sides] = z[13] z[edges] = -2. # force flow outwards from the tops of the guards z[hole_here] = -1. A_new = np.array([[[ 0., 1., 1., 1., 1., 1., 0., 0., 1., 1., 1., 1., 1., 0., 15., 5., 4., 3., 2., 1., 0., 0., 10., 4., 3., 2., 1., 0., 0., 1., 4., 3., 2., 1., 0., 0., 1., 1., 1., 1., 1., 0., 0., 1., 1., 1., 1., 1., 0., ]]]).flatten() depr_outlet_target = np.array([ XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, 14, 14, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, XX, ]).flatten() mg.add_field("node", "topographic__elevation", z, units="-") fr = FlowRouter(mg) lf = DepressionFinderAndRouter(mg) fr.route_flow() lf.map_depressions() assert mg.at_node["drainage_area"] == approx(A_new) assert lf.depression_outlet_map == approx(depr_outlet_target)